Friday, December 1, 2023

Even crypto mixing deserves a threshold

Many of you may not realize this, but in most parts of the developed world, banks automatically record and report our transactions to law enforcement. The logic behind this is that by giving up our personal data, we get more security, albeit at the cost of 1) losing our privacy, and 2) adding an extra layer of costly red tape into financial life.

It's a pragmatic compromise, and one hopes that the benefits outweigh the costs. The way that we've been balancing this compromise up till now is by using thresholds, so as to reduce the cost side of the equation. Below a certain dollar threshold (i.e. $10,000 for cash), transactions don't get reported. The folks making these sub-threshold transactions thus enjoy the dignity of not having their privacy invaded, nor do they add to the financial sector's administrative burden. However, they also don't contribute to the effort to improve security and safety.

Anyways, last month, the U.S. government announced a new anti-money laundering reporting requirement, one for crypto mixing. In doing so it broke with a long tradition of not including a threshold. That got my hackles up. Thresholds have always been key to balancing the costs and benefits of automatic reporting requirements.

In short, the government thinks that mixing of cryptocurrency is of primary money laundering concern. Any U.S. financial institution that knows, suspects, or has reason to suspect that a customer's incoming or outgoing crypto transaction, in any amount, involves the use of a mixer will have to flag it and send a report to the government. That report must include information like the customer's name, date of birth, address, and tax ID. 

I submitted the following comment on the proposed rule for crypto mixing. If you agree, feel free to copy it and add your own comment to the growing pile. 

Dear sir/madam,

Re: Proposal of Special Measure Regarding Convertible Virtual Currency Mixing, as a Class of Transactions of Primary Money Laundering Concern

Historically, all U.S. anti-money laundering recordkeeping and reporting requirements have been accompanied by a monetary threshold. The current proposal to impose recordkeeping and reporting requirements for crypto mixing is the sole exception. This should be fixed.

When Treasury Secretary Henry Morgenthau published an executive order to implement the U.S.'s first large cash transaction reporting regime all the way back in 1945, for instance, he established a $1,000 reporting requirement for transactions in which only bills in denominations over $50 were present. He also set a $10,000 reporting threshold when small and large denomination bills were involved in the transaction.

Morgenthau's thresholds remained in place through the 1950s and 1960s. They were eventually ratified in 1972 with the implementation of a $10,000 cash reporting threshold for the purposes of implementing the Bank Secrecy Act.

When suspicious activity reports were introduced in 1996, the government's initial proposal did not include a reporting threshold. But after receiving public comments, the government admitted that its first version of the rule would impose a "burden of reporting." In its final version it introduced a $5,000 threshold for filing a suspicious activity report, which remains to this day.

In addition to reporting thresholds for cash transactions and suspicious activity, the government has set a number of thresholds for recordkeeping requirements. For instance, financial institutions are required to keep a log of all cash purchases of monetary instruments between $3,000 and $10,000.

The government's long history of twinning reporting and recordkeeping requirements with thresholds is a pragmatic compromise. It balances law enforcement's need for information against the administrative burden imposed on the private sector as well the invasion of privacy imposed on civil society. It only seems fair and prudent to extend this pragmatic compromise to cryptocurrency mixing recordkeeping and reporting requirements, especially in light of the fact that, as FinCEN admits, there are "legitimate purposes" for mixing.

I would suggest a threshold of at least $10,000, which is in-line with the cash transaction reporting threshold.

JP Koning
Moneyness Blog

Tuesday, November 28, 2023

Are central banks too reliant on SWIFT for domestic payments?

Central bank settlement systems are the the tectonic plates of the payment system: they are vitally important to our lives, but we never see them in action. All of a nations' electronic payments are ultimately completed, or settled, on these systems. If they stop working, our financial lives go on pause, or at least regress to older forms of payment.

In this post I want to introduce readers to a crucial feature of these payments tectonic plates: their reliance for domestic settlement on SWIFTNet, a financial messaging network used by banks and other financial institutions to communicate payments information. Think of SWIFTNet as a WhatsApp for banks, but exclusive and very secure. 

This reliance  or over-reliance  is best exemplified by a recent decision by the European Central Bank. The Target2 settlement system has long been the bedrock layer of the European payments universe. All domestic payment ultimately get tied-off on the system. Since it was introduced in 2007, Target2 has been solely reliant on SWIFTNet for sending and receiving messages. 

When the European Central Bank replaced Target2 with T2 earlier this year, it modified the system to have two access points: it kept SWIFTNet but added a competing messaging network, SIAnet, to the mix. As one commentator triumphantly put it, "SWIFT’s monopoly for access to the T2/T2S system is broken."

SWIFTNet is owned by the Society for Worldwide Interbank Financial Telecommunication, or SWIFT, which is structured as a cooperative society under Belgian law and is owned and governed by its 11,000 or so member financial institutions. Whenever SWIFT gets mentioned in conversations, it tends to be associated with cross-border wire payments, for which its messaging network is dominant. However, for many jurisdictions, including Europe, SWIFT is also integral to making domestic payments. It's this little-known local reliance that I'm going to explore in this post.

The dilemma faced by central banks such as the European Central Bank is that SWIFTNet is an incredibly useful messaging network. It is ubiquitous: most banks already use it for cross-border payments. And so the path of least resistance for many central banks is to outsource a nation's domestic messaging requirements to SWIFT, too. However, this reliance exposes national infrastructure to SWIFTNet-related risks like foreign control, sanctions, snooping, and system outages.

Financial messaging 101

Before going further, we need to understand why financial messaging is important. For a single electronic payment to be completed, a set of databases owned by a number of financial institutions, usually banks, must engage in an intricate dance of credits and debits. To coordinate this dance, these banks need to communicate, and that's where a messaging network is crucial.

Say, for example, that Google needs to pay Apple $10 million. Google tells its banker at Wells Fargo to make the payment. Wells Fargo first updates its own database by debiting Google's balance by $10 million. The payment now has to hop over to Google Apple, which banks at Chase. For that to happen the payment flow must progress to the core of the U.S's payments system, the database owned by the Federal Reserve, the U.S.'s central bank.

Along with most other U.S. banks, Wells Fargo has an account at the Federal Reserve. It communicates to the central bank that it wants its balance to be debited by $10 million and the account of Chase to be credited by that amount. Once Chase's account at the Federal Reserve is updated, Chase gets a notification that it can finally credit Apple for $10 million. At that point Apple can finally spend the $10 million.

This entire process takes just a second or two. For this "dance of databases" to execute properly, the Federal Reserve, Chase, and Wells Fargo need to be connected to a communications network.

The sort of messaging network to which the central bank is connected, and the stewardship of that network, is thus crucial to the entire functioning of the economy.

Proprietary messaging networks or SWIFTNet? 

The Federal Reserve is somewhat unique among central banks in that it has built its own proprietary messaging network for banks. All of the 9,000 or so financial institutions that use the Federal Reserve settlement system, Fedwire, must connect to the Fed's proprietary messaging network to make Fedwire payments. To make international payments, however, U.S. banks must still communicate via SWIFTNet.  

Let's flesh the story out by trekking north of the border. Whereas the Federal Reserve has no reliance on SWIFTNet, Canada's core piece of domestic settlement infrastructure, Lynx, relies entirely on SWIFTNet for messaging.

For example, if Toronto Dominion Bank needs to make a $10 million to Scotiabank, it enters this order into SWIFTNet, upon which SWIFT forwards the message to Lynx, which updates each banks' accounts by $10 million and sends a confirmation back to SWIFTNet, which tells Scotiabank that the payment has settled.

For payments nerds, this network setup is called a Y-copy topology. The network looks like a "Y" because the originating bank message is relayed from the sending bank via SWIFTNet, the pivot at the center of the Y, down to the settlement system, and then back up via SWIFTNet to the recipient bank. It is illustrated below in the context of the UK's payment system, with the CHAPS settlement system instead of Lynx, but the idea is the same.

A Y-copy network topology for settling central bank payments in the UK [source]

The upshot is that the Federal Reserve controls the messaging apparatus on which its domestic settlement depends, whereas Canada outsources this to a cooperative on the other side of the ocean.

Many of the world's small and middle-sized central banks have adopted the same Y-copy approach as Canada. This list includes Australia, Singapore, New Zealand, Nigeria, UK, Sweden and South Africa. However, some members of this group are starting to have second thoughts about fusing themselves so completely to SWIFT.

Removing the single point of failure

The European Central Bank is at the vanguard of this group. Prior to 2023, the European Central Bank was in the same bucket as Canada, relying entirely on SWIFTNet to settle domestic transactions. 

With its upgraded T2 system, Europe doesn't go quite as far the Fed's model, which is to build its own bespoke messaging network. Rather, European banks now have the option of either sending messages to T2 using SWIFTNet, or they can use SIAnet, a competing network owned by Nexi, a publicly-traded corporation. SIAnet stands for Societa Interbancaria per l'Automazione, a network that originally connected Italian banks but has now gone pan-European.

The reason for this design switch is that European Central Bank desires "network-agnostic connectivity." This dual access model will make things more complex for the European Central Bank. If a commercial bank originates a SIAnet message, the central bank will have to translate this over to a SWIFT message if the recipient bank uses SWIFTNet. Nevertheless, the European Central Bank believes this dual structure will offer more choice to domestic banks.

The ECB also hints at the enhanced "information security" that this new setup will provide, without providing much detail. The UK's recent efforts to update its core settlement layer sheds some extra insights into what these security improvements might be. Right now, the UK's core settlement system, CHAPS, can only be accessed by SWIFTNet, much like in Canada, so that all domestic UK payments are SWIFT-reliant.

In its roadmap for updating CHAPS, the Bank of England is proposing to allow banks to access the system via either SWIFTNet or a second network, which doesn't yet exist. The idea is to enable "resilient connectivity" to the core settlement layer, especially in periods of "operational or market disruption." Should SWIFTNet go down there would be no way for financial institutions to communicate with CHAPS, and the entire domestic economy would grind to a halt. A second network removes the "single point of failure" by allowing banks to re-route messages to CHAPS.

The Bank of England also highlights the benefits of competition, which would reduce the costs of connectivity.

This sounds great, but there are tradeoffs. Using a a single network for both domestic and international payments is valuable to the private sector because it offers standardization and efficiencies in banks' processing. Adding a second option will also complicate things for the Bank of England, since it will have to design and build a system from scratch, much like the Fed did, which could be costly. Either that or it will have to find another private option, like the ECB did with SIAnet. This second network may not be as good as SWIFTNet which, despite worries about resiliency, has been incredibly successful.

When CHAPS went down earlier this year for a few hours, for instance, it wasn't SWIFT's fault, but the Bank of England's fault. The same goes for a full day outage in 2014. 

Comparing a V-shaped network topology to Y-Copy in an Australian context [source]

The type of settlement topology that the UK is proposing is known as "V-shaped," since all messages are sent directly to the central bank settlement system for processing via any of a number of messaging networks, and then back to the recipient bank. The difference between a V-shaped topology and Y-copy is visualized in the chart above in an Australian context, but the principles apply just as well to the UK.

Sanctions and "the SWIFT affair"

The decision to make domestic payments less dependent on SWIFTNet is much more easy to make for outlier nations like Russia. SWIFT is based in Belgium and is overseen by the Belgian central bank, along with the G-10 central banks: Banca d’Italia, Bank of Canada, Bank of England, Bank of Japan, Banque de France, De Nederlandsche Bank, Deutsche Bundesbank, European Central Bank, Sveriges Riksbank, Swiss National Bank, and the Federal Reserve. That put SWIFT governance far out of Russian control.

You can see why this could be a problem for Russia. Imagine that only way to settle domestic Russian payments was by communicating through SWIFTNet. If Russia was subsequently cut off from that network for violating international law, that would mean that all Russian domestic payments would suddenly cease to work. It would be a disaster.

Needless to say, the Central Bank of Russia has ensured that it doesn't depend on SWIFTNet for communications. It has its own domestic messaging network known as Sistema peredachi finansovykh soobscheniy, or System for Transfer of Financial Messages (SPFS), which was built in 2014 after the invasion of Crimea. Prior to then, it appears that "almost all" domestic Russian transactions passed through SWIFTNet  a dangerous proposition for a country about to face sanctions.

Mind you, while Russia has protected its domestic payments from SWIFTNet-related risk, it can't do the same for its international payments. SWIFTNet remains the dominant network for making a cross border wire. There is no network the Russians can create that will get around this.

I'm pretty sure that most larger developing states and/or rogue nations have long-since built independent domestic financial messaging systems to avoid SWIFTNet risk. I believe China has done so. Brazil has the National Financial System Network, or Rede do sistema financeiro nacional (RSFN). India also has its own system, the Structured Financial Messaging System (SFMS), built in 2001. India is even trying to export SFMS as a SWIFT competitor.

The Japanese were typically way ahead on this. The Bank of Japan built its messaging network, the Zengin Data Telecommunication System, back in 1973, several years before SWIFT was founded.

The last SWIFTNet risk is snooping risk. This gets us into the so-called SWIFT affair. After 9/11, the U.S. intelligence agencies were able to pry open SWIFT through secret broad administrative subpoenas. They had the jurisdiction to do so because one of SWIFT's two main data centres was located in the U.S.

To ensure data integrity, SWIFT had been mirroring European data held in its data centre in Belgium at its U.S. site. That effectively gave U.S. intelligence access to not only SWIFT's U.S. payments information, but  also information on foreign payments sourced from Europe or directed to Europe. Worse, it also provided spooks with data on domestic European payments. Recall that the European Central Bank's Target2 settlement system, which settles all digital domestic payments in Europe, was entirely reliant on SWIFTNet for communications.

When the U.S.'s snooping arrangement was made public by the New York Times in 2006, it caused a huge controversy in Europe. SWIFT tried to placate Europe by building a third data warehouse in Switzerland to house Europe's back-up data. But the precedent was set: SWIFT is not 100% trustworthy. And that may be part of the reason why the European Central Bank chose to downgrade its reliance on SWIFTNet when it introduced its new system, and is surely why other nations want to entirely hive their domestic systems off from it.


In sum, central banks face a host of complicated decisions in how to bolt on messaging capabilities to their key settlement systems. SWIFTNet is a top notch network. However, too much SWIFT-related risk may be perceived as having negative implications for national security. For large nations with extensive banking industries, building a proprietary domestic messaging alternative seems to be the preferred option. It also seems to be the default choice for rogue states like Russia.

Another alternative is to fallback on using multiple independent networks for access, of which one is SWIFTNet, and thus mitigating exposure to SWIFT-related problems. This is the approach taken by Europe and the UK.

For smaller nations that comply with the global consensus, like Canada, the calculus is different. Building an alternative communications network is likely to be costly. The risk of sanctions and censorship are negligible while the benefits of using a high-quality ubiquitous network for both domestic and foreign payments messaging are significant. Given these factors, it may be worthwhile to bear all SWIFT-related risks and adopt the Y-copy model.

Monday, November 20, 2023

Is it legal to mix cash in a jar?

Chris Blec asks the following question:

Source: Twitter

Chris's premise is that it is "not illegal" for him to get together with a bunch of strangers to mix cash. But that's not quite right. It can be legal. It can also be illegal. To determine which it is, we need to understand the motivations of Chris and the other ten strangers. Why are they getting together to mix in the first place? Alas, Chris doesn't mention this in his tweet.

There are certainly all sorts of perfectly legal albeit quirky reasons to mix cash. We could imagine that Chris and ten other strangers are waiting for the bus, and to decide who goes first, they all put a $20 note into a jar, remembering their respective serial number. After the notes have been mixed, one note is taken out and whoever it belongs to wins. The notes are then given back. Nothing wrong with that.

On the other hand, if Chris and the other 10 strangers are mixing their cash because they want to conceal the source, then they need to be careful. They've taken one step down the path to engaging in money laundering.

The U.S. has several money laundering statutes. Below is part of one of the most contravened ones: 

Source: LII

As you can see, one of the key triggers for a money laundering conviction is making transactions that are designed to "conceal or disguise."

Even if Chris and the other strangers' motivations for mixing is to conceal the origins of their cash, that's not necessarily illegal. Before they reach the point at which they can be accused of having crossed the line over to money laundering, at least one of the strangers needs to contribute banknotes to the jar that are the "proceeds of specified unlawful activity." For argument's sake, let's say that one of the strangers contributes cash that they've earned from contract killing. Chris and the other 10 strangers are now a step closer to a potential money laundering indictment. 

Only one last criteria is lacking. Chris and the other strangers must participate in a "knowing" way. They must be aware that the property involved is criminally-derived. The most obvious example would be if one of the 10 strangers were to loudly announce just prior to putting their notes in the jar that the notes come from contract killing, and everyone hears this yet still participates. 

At this point, the three triggers have been met. Chris and the other strangers have acted in 1) a knowing way 2) to conceal 3) the actual proceeds of unlawful activity.

The state of "knowing" needn't be established in such an explicit fashion as the criminal announcing it loudly. For instance, even if the criminal says nothing, but Chris and the other participants suspect the possibility that dirty money is entering the jar, but they don't do due diligence, then they could be found guilty of money laundering. To demonstrate that they aren't knowing participants, Chris and the other strangers may have to take proactive measures, say like checking ID.

So Chris is right to say that mixing cash in a jar can be legal, but he incorrectly omits to say that it can also be illegal.

Having fleshed out Chris's premise, what about his conclusion? Can the jar-of-cash thought experiment teach us about the legality of crypto mixing methods such as custodial mixers, Tornado Cash, or CoinJoin? I'll let the readers work that one out on their own.

Thursday, November 16, 2023

Kraken v Kraken, or how to protect the public from crypto exchange failures

It's been a full year since FTX International and FTX-US collapsed, and the shocking thing is  there is still no regulated crypto venue in the U.S.! You'd think some lessons would have been learnt.

To best protect the public from the Sam Bankman-Frieds of the world, what the U.S. requires is securities-level oversight of crypto exchanges. Exchanges like Coinbase and Kraken are offering the sorts of investment services to the public that the U.S.'s main securities regulator, the SEC, is ideally positioned to regulate, such as trading, margin, custody, and market making. But one year after FTX's collapse, there doesn't appear to be a single SEC-regulated exchange.

The exchanges blame this on the SEC's lack of clarity. The SEC blames this on exchanges refusing to come in and register. God knows who's telling the truth.

Whatever the case, this intransigence only increases the odds that there'll be another U.S. crypto exchange collapse in the next few years, one that appropriate regulation could have otherwise prevented, or at least sheltered investors from the fallout.

What sort of protections am I talking about? After Canada suffered through the collapse of QuadrigaCX a few years back, and a bunch of Canadians like myself lost money, crypto exchanges were brought under the auspices of our existing securities regulatory framework, with a few nips and tucks to the rules to make them fit. This has led to a lot of changes that make Canadian crypto customers safer. I'm going to share the best example in this blog post.

Kraken is a well-known crypto exchange that serves both American and Canadian customers. However, if you're an American customers who uses Kraken, you possess a very different sort of asset than Canadian customers do.

Let's look at the fine print of the U.S. platform:

Source: Kraken's terms of service for U.S. customers

So in the U.S., asset are held "by us for you." That is, Kraken itself is doing the holding, safekeeping, or providing custody of crypto for its customers.

A key observation I want to make here is how fundamentally different this is from how standard regulated marketplaces function like the NASDAQ or the Toronto Stock Exchange. Traditional marketplaces offer a venue to trade assets, but they don't offer custody. If they tried to introduce this, their regulator would very quickly say no. I'll explain why below.

But first, let's head over to Kraken Canada. Once again, here's the fine-print:

Source: Kraken's terms of service for Canadian customers

What the underlined wording says is that if you're a Canadian, Kraken does not hold your crypto for you. That's very different from the U.S. Instead, Kraken says that customer crypto is deemed to be "custodial assets" and delegates your crypto to a "designated trust account at a Crypto Custodian." Bingo. There's the separation of trading from custody that I was talking about earlier, which aligns with standard practices for marketplaces.

Scan further through the fine print and you learn who that crypto custodian is: Anchorage Digital Bank:

Source: Kraken's terms of service for Canadian customers

Who is Anchorage? Anchorage is a federally-charted trust that is overseen by the Office of the Comptroller of the Currency, one of the key U.S. federal banking regulators. So if you hold some coins on Kraken, and you are an American, you own what is essentially a Kraken IOU, and you have to trust Kraken, but if you are a Canadian, you're effectively putting most of your trust in a federally charted financial institution. That's pretty stern stuff. 

Canadian securities regulators require all crypto exchanges operating in Canada to delegate at least 80% of customer crypto to a third-party custodian. (The other 20% can be held in a hot wallet for liquidity purposes.) Kraken doesn't appear to be doing this for its American customers, and that's because there's no U.S. regulator prompting it to do so. On top of requiring this separation, Canadian regulators stipulate that third-party custodians must be qualified. That is, they can't just walk in off the street. The custodian has to meet the regulator's standards, which requires having a Systems and Organization Controls (SOC) designation, and a bunch of other stuff too.

You can probably see by now that if you're a customer of Kraken, it's better to be Canadian than American, for the following reasons:

Anchorage is a federally-regulated financial institution and subject to strict oversight. Kraken U.S. is for the most part unregulated. No one is peering over its shoulder to check whether it is doing a good job safekeeping your coins.
Kraken has its fingers in a lot of different businesses, but Anchorage specializes on custody, and so it's probably better at the task.
Anchorage is independent from Kraken. This separation mitigates the risk of loss, theft, or misuse of assets by Kraken management. This is particularly salient in Kraken's case because it engages in many other business activities, such as trading or market-making, and these pose potential conflicts of interest.

In the future, one hopes that Kraken's U.S. exchange provides the same level of customer protection as Kraken's Canadian platform. But that's only go to happen if and when the SEC dictates a fundamental separation of crypto trading from custody, and whether U.S. crypto exchanges actually listen to the SEC.

Addendum (Nov 21): Talk about good timing. The SEC just announced that it is suing Kraken's U.S. entity for, among many other things, failing to segregate customer crypto assets and dollar balances. Segregation is different than third party custody. It basically means that customer property is kept separate from corporate property. This helps to prevent double-dipping. An exchange can make use of a third party custodian, for instance, but not segregate those funds into corporate and customer buckets. The combination of segregation and third-party custody is optimal.

By contrast, Canadian securities law clearly specifies that Kraken and any other Canadian exchange must already be segregating customers crypto and funds from corporate funds. In the regulators' own words, exchanges must keep customer property "separate and apart from its own property." This is in addition to the requirement that they use a third party custodian to store customer crypto and fiat. We can verify by reading Kraken's undertaking with Canadian regulators, in which it promises that it will be keeping customer crypto "separate and apart from its own assets."

Segregation is just one other low-hanging bit of customer protection that U.S. crypto exchanges should already have implemented, but probably won't until prodded by the government.

The next section is for pedants only, of which I'm embarrassed to be, which is why I'm putting this in very small print:

Kraken owns a U.S. bank. How does this fit into the story? Here are the details: In the U.S., the Kraken crypto exchange is really just the trade name for Payward Ventures. Payward Ventures is in turn a subsidiary of Payward, Inc. Payward, Inc has another subsidiary, Payward Financial, Inc, that owns a state-charted bank -- Kraken Bank. Notably, when you sign up as a customer of the Kraken crypto exchange, you are entering into a relationship with Payward Ventures, not Payward Financial. There is no indication in the exchange's terms of service that Kraken Bank is in any way involved in custody. Which seems... odd? Why wouldn't Kraken use its bank for custody?

When Kraken first applied to do business in Canada earlier this year, it said it wanted to use Kraken Bank as its custodian. Given that Kraken is in fact using Anchorage Bank as we speak, I suspect that Canadian regulators told Kraken: "Hey, guys. Kraken Bank is not sufficiently independent, you're going to have to use a third-party." And I suspect they were right about this.

Meanwhile, what about Canada's most popular exchange, Coinbase? Coinbase's Canadian terms of service doesn't indicate that it is using a qualified custodian. Customer assets are "held by the Coinbase Group for your benefit." Yeah, that's not going to fly with the regulators. I suspect that within a few months you're going to see it using a third-party like Anchorage, or its just going to leave Canada.

Tuesday, November 14, 2023

In praise of anti-money laundering thresholds

Two seemingly separate stories, a crypto and a banking story, have a common thread in anti-money laundering thresholds.

In the first story, the New York Times shows how regular folks are increasingly losing their bank accounts because their bank perceives them to be engaging in risky behaviour. In the second, the U.S. government has proposed an expansive new rule that would require financial institutions to report all customers who use cryptocurrency mixers to the government.

Anti-money laundering thresholds underpin what I'll call the Pragmatic Compromise between the government and citizens, albeit a tenuous compromise, for reasons I'll explain.

The U.S. government and its various law enforcement agencies have the ability to get full access to bank records for the purposes of fighting crime. They can do so directly, that is, without having to proceed through the standard process of convincing a judge to approve a warrant. This is an incredible amount of power to have. To counterbalance this, a compromise of sorts has been agreed to that limits the government's access to bank records. A number of key financial thresholds have been established, below which transactions are protected from surveillance.

The most well-known method the government has for accessing your personal financial information is the requirement that banks submit currency transaction reports, or CTRs (see below), every time someone withdraws or deposits paper money. Banks don't submit a report for all cash transactions. The threshold for submitting is set at $10,000. So if you withdraw $9,999, your name and address won't be reported to the government. If you withdraw $10,001, you'll lose the threshold's protection and will be reported.

The modern U.S. currency transaction report [source]

The U.S. has a long history of providing direct government access to banking records. The practice began in 1945 when Treasury Secretary Henry Morgenthau Jr, invoking war-time powers, issued an executive order (see below) instructing banks to begin reporting currency deposits and withdrawals made by the public. Known as TCR-1 forms, these reports were to be forwarded every month to the government with information about the cash amounts involved and the identification of the individual making the transaction.

Morgenthau's reporting requirement was motivated by the desire to stamp out black marketeering, writes Paul Camacho, which had emerged as a way to evade war-time rationing programs. But even though the war soon ended and rationing ceased, the practice of cash reporting continued through the 1950s and 1960s, albeit on what must have been legally dubious grounds now that it was peacetime.

Henry Morgenthau's 1945 executive order on currency reporting [source]

In 1970, the necessary legal formalization to justify the reporting of bank information was provided when Congress passed the Bank Secrecy Act, which encoded directly into law the requirement that banks record and report cash transactions, as well as legislating a set of extra recordkeeping and reporting requirements. Over the years, additional recordkeeping and reporting standards were added by Congress to the Bank Secrecy Act, including the 1994 requirement that banks screen for "suspicious" transactions and report them to the government by submitting a suspicious activity report, or SAR (see below).

While there's certainly a law & order case to be made for providing governments with direct (i.e. warrantless) access to financial records, there's a pretty clear set of reasons why society should want to limit this power. Allow too much access and banks will inevitably get bogged down in the expensive bureaucracy of filing reports, which can lead to accessibility problems as the accounts of customers deemed to be a compliance nuissance are terminated—especially the ones who tend to make riskier but legal transactions. There's also the crucial question of the public's right not to be be snooped on, especially without a warrant and probable cause.

Suspicious Activity Report form, introduced in 1995

The first legal challenges to the government's direct access to bank records only came in the mid 1970s. But after hearing these cases (California Bankers Association v. Shultz and United States v. Miller), the Supreme Court allowed the entire data collection apparatus to remain intact. The majority ruled that there was no expectation of privacy in bank records, and that the recordkeeping and reporting requirements imposed by the Bank Secrecy Act do not violate bank customers' constitutional rights.

With the public having no constitutional protections against direct government access to bank records, the lone remaining counterbalance is the various anti-money laundering thresholds.

Which gets us back to the New York Times article (we'll touch on the cryptocurrency further down). The reasons for the growing debanking problem that the Times article highlights is complicated, but I'd suggest that one driver is a steady deterioration of two key reporting thresholds at the heart of the Pragmatic Compromise.

The original $10,000 cash reporting threshold was set back in 1945 by Henry Morgenthau, a level that was ratified in 1972 after the passage of the Bank Secrecy Act. This level has never been adjusted. (Morgenthau also set a second and lower $1,000 threshohold, but this only applied when banknotes in denominations of $50 or higher were involved).

Alas, inflation has been steadily eating into each thresholds' real value. When the Bank Secrecy Act was passed, $10,000 was worth $75,000 in today's dollars. In Morgenthau's time it was equal to $173,000. Either way, when the data collection apparatus was first established and the Pragmatic Compromise reached, most people's day-to-day cash withdrawals and deposits would have been sheltered from reported requirements. With the passage of time and inflation, a much wider swathe of civilian cash transactions have lost the protection offered by Morgenthau's $10,000 threshold. That means more snooping. It also means more debanking. Rather than absorbing the growing compliance costs of having "risky" cash-reliant customers, banks are closing accounts.  

As for suspicious activity reports, when they were first legislated in 1994 the government subjected them to a $5,000 threshold, which is equal to around $10,000 today. With inflation having effectively destroyed half of the value of the threshold, more and more regular transactions are falling under suspicion. As the Times points out, suspicious customers don't make for good customers: "Multiple SARs often — though not always — lead to a customer’s eviction."

The Pragmatic Compromise that society came to decades ago is being poorly administered. To restore it, what is needed is a reasonably-sized one-time "catching up" of the various thresholds to account for at least part of the inflation that has occurred over the years, and then periodic adjustments to these levels each year to account for subsequent inflation. Maybe that'll solve some of the problems brought to light by the Times.

Now let's turn to the crypto story. In addition to the two classic anti-money laundering reporting requirements, CTRs and SARs, the U.S. government is now proposing a third reporting requirement: one for crypto mixing.

Last month, the government announced that it deems mixing of cryptocurrency to be of primary money laundering concern. Any U.S. financial institution that knows, suspects, or has reason to suspect that a customer's incoming or outgoing crypto transaction involves the use of a mixer will have to flag it and send a report to the government. That report must include information like the customer's name, date of birth, address, and tax ID.

The US Treasury's proposed rule for treating crypto mixing as a primary money laundering concern [link]

Notably, there are no thresholds to this proposed reporting requirement. Any customer crypto transfer with even just a whiff of mixing must be reported by financial institutions to the government.

In its proposal (it isn't final, yet) the government grants that there are "legitimate purposes" for mixing. What are they? I think the popular view of crypto is that it is anonymous, but this isn't quite right. Every bitcoin or ether transaction gets recorded on transparent databases. This makes them trackable by anyone. In some respects, crypto may be the least privacy-friendly financial medium ever created. Mixing your coins in a jumble along with other's coins is one of the ways to free oneself from this all-seeing eye, both for criminals who have stolen coins and regular folks who don't want their financial lives displayed for all to see.

Given that there can be licit reasons for mixing, and putting this in the context of the decades-old Pragmatic Compromise, precedent would seem to suggest that the government should include a reasonably-sized threshold for reporting crypto mixing. If this was set at, say, $10,000 (and then adjusted yearly for inflation), then a customer could in theory mix $8,000 worth of bitcoins and then deposit them at an exchange, and that exchange would not need to report the transaction and the person who made it. Go above $10,000, and the customer will end up in a government computer.

Without a reasonable threshold, the same phenomenon that the Times documents with respect to bank customers will happen to crypto users. A wave of deplatforming will hit as financial institutions close the accounts of any customer that betrays even a hint of risky activity, much of which might only appear to be mixing when it isn't.

The government has the capacity for changing its initial stance on thresholds. If you go back to the early 1970s when the Bank Secrecy Act's thresholds were set by executive order, the government initially proposed a $5,000 threshold. After the public provided comments and grievances were aired, the final rule compromised and pushed it up to $10,000.

Likewise for SARs.

The government's original 1995 proposed rulemaking for suspicious activity reporting included no thresholds. So even a $10 or $20 suspicious payment would have been reported to the government. In response to public push-back, the government admitted that its first version of the rule would impose a "burden of reporting," and in its final version it introduced a $5,000 threshold for filing a SAR, which the U.S. has to this day.

The government also granted in its 1995 SAR decision that the adoption of a threshold was intended to "conform the treatment of money laundering and related transactions to that of other situations in which reporting is required." This seems to me to be a pretty clear admission of the Pragmatic Compromise; that reasonably sized dollar thresholds are a standard element of any anti-money laundering reporting requirement. I don't see any reason why the government's 1995 olive branch for suspicious activity reporting should not be extended to crypto mixing.

Wednesday, November 8, 2023

How would a cash-only central bank conduct monetary policy?

What role do changes in the supply of banknotes play in contributing to a central bank's ability to carry out monetary policy? Put differently, to what degree does "printing," or creating new physical currency and issuing it into the economy, contribute to generating a central bank's desired inflation rate of 2-3%?

In a recent blog post at Econlog, Scott Sumner suggests that printing physical cash and "forcing" or "injecting" it into the economy has been an important part of central banks hitting their inflation targets, albeit less so now than in times past. I'm not so sure.

Having imbibed Scott's blog posts for more than a decade, I think I'm 99% on the same page as he is when it comes to thinking about monetary policy. We both agree that a central bank must either reduce the interest rate that it pays on the monetary base, or inject more monetary base into the economy, in order to push up prices. Using either of these two methods, the central bank sets off a hot potato effect in which a long chain of market participants do their best to unload their excess money balances, a process that only comes to an end when all prices have risen to a new and higher equilibrium such that no one feels any additional urge to spend away their extra money. Scott once described the hot potato effect as the the "sine qua non of monetary economics."

The monetary base is comprised of two central bank financial instruments: physical banknotes (a.k.a. cash) and digital clearing balances, sometimes known as reserves, a type of money used by banks.

Reading through Scott's post, I think the one spot where we may disagree is on the relative role played by the two types of base money in the hot potato process.

For my part, I don't think that cash has ever had much of a primary role to play in setting off a hot potato effect. All of the initial uumph necessary for driving prices towards target has typically been provided by reserves, either via a change in the interest rate on reserves or a change in their quantity. Once that initial uumph has been delivered, a whole host of other money types  physical currency, bank deposits, checks, money market funds, and PayPal balances  helps convey the forces originally unleashed by reserves to all corners of the economy.

An example of the hot potato effect in action may help illustrate.

Let's start with a central bank that needs to push inflation up to target. It reduces the interest rates on reserves. The first reaction to lower rates is a flight out of reserves into other assets, say shares.

As a result, share prices quickly rise. Those existing shareholders who realized their gains by selling at the new and higher price now find themselves with a hot potato on their hands; they have too many monetary balances in their possession and not enough non-monetary things.

Some of these ex-shareholders may choose to spend their excess deposits to go on, say, a vacation. As a result, airline ticket prices rise. Others transfer their extra money to their PayPal account in order to send it to friends and family, who may in turn make purchases, pushing up the prices of whatever they buy. Another group of ex-shareholders decides to buy used cars. They withdraw banknotes, their banks in turn asking the Fed to print new banknotes and ship them over. Used car prices rise.

The point is, the initial uumph is delivered by the change in reserves, and this gets conveyed to all prices by a daisy-chain of spenders offloading an array of different types of excess money.

In this story, note that cash isn't being actively "injected" into the economy by central banks, nor by commercial banks. Rather, people are choosing to draw cash out as their preferred method for getting rid of unwanted money, in response to a set of forces initiated by reserves. Reserves are the central bank's lever for change; cash is merely responsive.

Consider too that in a world where cash no longer exists, and has been replaced by digital payments options, monetary policy is still effective. In this world, the response to a reduction in the interest rate on reserves gets conducted to all the economy's nooks and crannies via non-cash types of monies, like fintech balances and bank deposits. (I'd be curious to hear if Scott is of the same opinion about monetary policy in a world without cash.)

Here's an interesting thought experiment. Would it be possible to redesign cash and reserves in such a way that cash takes over the initiatory role in monetary policy from reserves? That is, can we turn cash into the active part of the monetary base, the one that drives changes in monetary policy, and relegate reserves to the passive role?  

One step we could take is to pay interest on cash. This may sound odd, but it's possible to do so by setting up a serial note lottery to pay, say, 3% per year to holders of cash. (I wrote about this idea here and here.) Simultaneously, reserves would be rendered less important by no longer paying interest on them.

Now when a central bank needs to raise consumer prices in order to hit its targets, it reduces the interest rate on cash from 3% to 2%. This ignites the hot potato process as the entire economy suddenly tries to offload its unwanted $20 and $100 bills, which at 2% just aren't as lucrative as before.

Another change we could enact would be to modify the mechanism by which central banks inject base money into the economy. As it stands now, central banks inject base money by purchasing assets with new reserves. Since reserves are digital, they are a lot more convenient for making billion dollar asset purchases than physical money. These extra reserves become hot potatoes in the hands of asset sellers, which sets off the process of price adjustments described in previous paragraphs. If central banks were to buy assets with cash rather than reserves, that would put cash in the driver's seat, albeit at the expense of convenience.

It's an interesting thought experiment, but in the end I don't think it's very helpful to get bogged down over which type of base money has more monetary significance. As Scott says, the key point is that the central bank controls the price level via its control over base money in general. They can raise prices by either adding to the supply of base money, or by reducing the demand for base money with a cut in the interest rate paid on reserves. "It's basic supply and demand, nothing more." 

Wednesday, October 18, 2023

Crypto adoption in America

Source: America Loves Crypto

You may have recently come across the America Loves Crypto marketing campaign, sponsored by Coinbase, the U.S.'s largest crypto exchange. In an attempt to promote the voting power of crypto owners, the website makes the claim that 52 million American adults currently hold crypto, which constitutes 20% of the U.S. adult population. If true, that's a massive voting block.

As the source for its 20% statistic, Coinbase cites an online survey of 2,202 adults that it commissioned from Morning Consult last February, which among other questions queried respondents for how much crypto they currently hold.

If you've been following other sources for cryptocurrency adoption data, Coinbase's 20% statistic seems... questionable?

To see why, let's dig into U.S. crypto adoption data. What follows is a quick rundown of what the best surveys have had to say about Americans crypto ownership. Later on in the article I'll get into who they are, how much they own, and why they hold it.

1) The granddaddy of all U.S. payments surveys is the Federal Reserve's Survey and Diary of Consumer Payment Choice, or SDCPC. The SDCPC is a long-running data collection effort that tries to paint a comprehensive picture of U.S. consumers' payment preferences and behavior. It fortuitously began to incorporate crypto-related questions way back in 2014, although crypto is just a tiny portion of the incredible amount of data collected by the SDCPC.

The Fed's SDCPC is run through the University of Southern California's Understanding America Study panel. In its 2022 iteration the SDCPC polled more than 4,761 participants. Notably, the SDCPC includes both a survey and a 3-day diary portion. Diaries are more labor intensive to administer than surveys, but provide better information since they minimize recall bias.

The SDCPC found that 9.6% of American adults owned cryptocurrency in 2022, up from 9.1% in 2021, and far higher than the 0.6% it polled back in 2015. However, that's far less than Coinbase's 20% claim. Only one of these numbers can be right. Which one is it?

The SDCPC's historical findings are in the table below.

U.S. cryptocurrency adoptions rates. Source: Federal Reserve's 2022 Survey & Diary of Consumer Payment Choice

2) The Federal Reserve publishes another survey that also sheds light on U.S. crypto adoption. The Fed's annual Survey of Household Economics and Decisionmaking (SHED) examines the financial lives of American adults and their families, and is therefore more general than the Fed's SDCPC, which is focused exclusively on payments.

The Fed's SHED is administered by Ipsos using Ipsos's KnowlegePanel panel. In 2022, 11,667 participants completed the SHED.

The SHED only began to include questions about crypto in 2021. It finds that 10% of Americans used cryptocurrency during 2022, where "using" is defined as buying, holding, or making a payment or transfer with crypto. This amount was down from 12% in 2021 (see table below). The number of Americans who held cryptocurrency as an investment, a narrower definition than "using", fell to 8% in 2022 from 10%.

Source: Federal Reserve's 2022 SHED

The SHED's 8-10% number neatly confirms the SDCPC's 9.6% finding while disaffirming Coinbase's 20% statistic.

3) The next decent source for crypto adoption data is tabulated by a group of four economics and finance researchers using quarterly surveys of the Nielsen Homescan panel, comprised of 80,000 households. With response rates of 20-25% per survey, that represents data from 15,000 to 25,000 respondents.

Weber, Candia, Coibion, and Gorodnichenko (Weber et al) found that at the end of 2022, the fraction of all households owning crypto had risen to 12%. The black dotted line in the chart below shows how ownership rates have changed over time.

Source: Weber et al (2023) using Nielsen Homescan Panel data

4) The fourth in our survey of crypto surveys was conducted by the Pew Research Center using its American Trends Panel. In a March 2023 survey of 10,701 panelists, Pew found that 17% of American adults have "ever invested in, traded, or used" a cryptocurrency. This is a very broad category, and would presumably include someone who casually bought $25 worth of bitcoin back in 2015 and sold it three days later, and has never touched it again.

Drilling in further, of the 17% who have ever owned or used crypto, 69% report that they currently hold some, which works out to a 2023 ownership rate of 11-12% among American adults. That's not too far off the two Fed surveys and Weber et al, but significantly different from the Coinbase result.

5) A fifth source of data comes from Canada, which serves as a decent cross-check against U.S. data given that both countries are quite similar in terms of culture and geography. Of the two key Canadian surveys, the first is the Bank of Canada's long-running Bitcoin Omnibus Survey (BTCOS), administered by Ipsos, which amalgamates participants from three different panels.

The 2022 BTCOS polled 1,997 Canadians and found an ownership rate of 10%, down from 13% the year before (see chart below). This constitutes a lower bound to ownership rates since it only includes bitcoin owners. The 2022 BTCOS also finds that 3.5% of Canadians own dogecoin and 4% own ether. However, it's not possible to add these amounts to the 10% bitcoin ownership number since many respondents own multiple types of crypto. 

Source: Bank of Canada 2022 Bitcoin Omnibus Survey

The second Canadian survey of note was carried out by the Ontario Securities Commission in 2022 to explore Canadian attitudes towards crypto assets. Conducted with Ipsos, the survey polled 2,360 Canadians in early 2022. It found that 13% of Canadians currently own any type of cryptocurrency, including crypto ETFs, which are legal in Canada but illegal in the US.


So there you have it. The two Federal Reserve surveys put crypto ownership at 9.6% and 8-10% respectively in 2022, while Weber et al peg it at 12% using Nielsen Homescan data. Pew has American crypto ownership at 11-12% by early 2023. And in Canada, the Bank of Canada calculates bitcoin ownership to be at 10% by the end of 2022, while the Ontario Securities Commission pegs total crypto ownership at 13% in early 2022, before much of crypto imploded.

Given this range of data, the 20% adoption rate that Coinbase's Morning Consult survey trots out is a glaring outlier and probably deserves to be thrown out. The American crypto owner is a potentially sizeable voting group, but not as big as Coinbase would like us to believe.

For what it's worth, I've found a few other odd things with Coinbase's Morning Consult survey that further adds to my skepticism. Morning Consult reports that 8% of respondents currently own a cryptocurrency called USDC, down from 10% the quarter before (see here for more). The survey also says that 5% currently own Tether. USDC and Tether are stablecoins. To anyone who follows crypto closely, the idea that 1 in 10 Americans own any particular stablecoin is absurd. Given that Morning Consult has surely got this wrong, perhaps through a sampling error, that makes you wonder about the quality of their overall work.

However, even if we ignore Coinbase's Morning Consult survey, the 9.6% adoption rate found in the bellwether SDCPC is still breathtakingly high. In just fifteen years, crypto has gone from a strange niche product to something that is being held by tens of millions of Americans.

What additional facts do we know about America's crypto owners?

Size of holdings

Going through the SDCPC data, the majority of Americans who own crypto only have a little bit of the stuff. Out of all U.S. crypto owners surveyed, 45% owned just $0-200 worth of crypto in 2022. This is illustrated in the chart below. The median quantity of crypto held was $312. Given such small amounts, I doubt that these crypto owners qualify as durable crypto adopters, as opposed to folks who've jumped on the bandwagon when they saw a Superbowl ad from Coinbase, bought some dogecoin, and have since stopped paying any attention.

The SDCPC data suggests that 1 in 4 crypto owners are what I call crypto fundamentalists, holding more than $2,000 worth of crypto. Given that 90% of Americans don't hold any crypto at all, two in every 100 Americans qualifies as a crypto fundamentalist.

This skewed distribution of crypto ownership is confirmed by survey results from Weber et al and the Nielsen Homescan panel. An outlier group of hard-core owners, representing around 8% of all crypto owners, allocates their entire portfolio to crypto (see chart below). By far the largest group of crypto owners is comprised of small dabblers who put just 0-5% of their portfolios into crypto.

Source: Weber et al (2023) using Nielsen Homescan Panel data

Reasons for ownership

Why do Americans own cryptocurrency? Despite being labelled as "currencies," cryptocurrencies are not generally used as a medium for making payments. Price appreciation is the dominant motivation for owning them.

When the SDCPC survey queried participants in 2022 for their "primary reason for owning virtual currency," the most popular answer (at 67%) was investment (see chart below, orange rows). The second most popular reason (21%) was "I am interested in new technologies." It was rare for respondents to list any sort of payments-related use case as their primary reason for ownership. As for lack of trust in banks, the government, or the dollar  all of which are common cryptocurrency themes these were rarely mentioned in 2022 as a primary reason for ownership.

Interestingly, Americans crypto owners were not always so obsessed with price appreciation. In 2014, the SDCPC found that American crypto currency owners tended to offer a much broader set of motivations for owning crypto, including lack of trust, cross-border payments, and to make purchases of goods and services (see chart above, blue rows).

The dominance of investment as the motivating reason for owning crypto is echoed in Weber et al's analysis of Nielsen Homescan survey data. Respondents were allowed to give multiple reasons for owning cryptocurrency, the most popular reason (see chart below) being to take advantage of "expected increase in value." The desire to use cryptocurrencies for international transfers was almost nonexistent, as was the desire to be "independent of banks."

The Fed's SHED survey isolates the same pattern as the other two surveys (see table below). Of the 10% of Americans who report using cryptocurrency in 2022, most do so as an investment. One small difference is that the SHED reports that around 2% of all survey participants used crypto to send money to friends of family in 2022. This suggests the transactional motive, while not primary, may be somewhat more prevalent than the previous two surveys suggest.

Source: Federal Reserve's 2022 SHED

Types of cryptocurrency held

What sorts of crypto were popular with Americans? Using Nielsen Homescan data, Weber et al found that of the 11% of survey respondents who are crypto owners, 70% held bitcoin while just over 40% held ether and dogecoin respectively.

This distribution is echoed in the Fed's 2022 SDCPC. Almost 65% of all crypto owners surveyed held bitcoin (the data is here), making it the most popular type of cryptocurrency. Meanwhile, 44.8% held ether and 38% held dogecoin, a coin that was originally created as a joke in 2013. That works out to around 4-5% of all Americans who are in on the joke.

Crypto bros

There's a reason that people throw around the term "crypto bro." Without exception, all of the U.S. surveys find that cryptocurrency owners tend to be young, male, and have a high income. 

The same goes for Canada, where in 2022 there were three male bitcoin owners for every female. The Bank of Canada has also regularly tested bitcoin owners for their degree of financial literacy using the Big Three questions method, and finds that they tend to have lower financial literacy than non-bitcoin owners.

Interestingly, in addition to isolating the crypto bro pattern  the tendency for crypto owners to be young, male, and wealthy  both the Pew survey and the Fed's SHED (see table below) find that U.S. crypto owners are more likely to be Asian, followed by Black and Hispanic, and least likely to be White.

The Fed's SHED survey dug deeper into usage by demographics, and found that while "investment" remains the driving case for owning crypto, and that the rich use the stuff proportionally more than the poor, certain demographic groups tend to rely on it more than others for making transactions. Specifically, the SHED finds that among low income families, 5% report an investment motivation for holding crypto while 4% report a transactional motivation.

Source: Federal Reserve's 2022 SHED

All of this data suggests to me the existence of three American crypto archetypes. 

The most dominant crypto archetype is the young wealthy male crypto dabbler, most likely non-white, who holds a few hundred bucks worth of doge or some other coin in order to gamble on prices going up. Coinbase's America Loves Crypto campaign makes the claim that "the crypto owner is a critical voter," but I suspect this doesn't hold for the dominant dabbler archetype, who probably doesn't have much attachment to their casual $50 bet on doge or litecoin or bitcoin, and thus can't be assembled into a voting block for crypto-related cause.

Another archetype is the much rarer crypto fundamentalist, a young male who has committed most of his savings to crypto. I suspect these are the types I encounter on Twitter, who evangelize and debate crypto to anyone who'll listen. This may be a small group, but they are also the most likely to vote for crypto-related causes.

Lastly, there seems to be a very small group of low-income people who are using crypto for actual transactions, the original use-case that Satoshi Nakamoto intended when he introduced the first cryptocurrency back in 2008.

Friday, October 13, 2023

Inflation as a tax

Last week I explored how Henry VIII resorted to coin debasement as a way to raise revenues in order to fight his wars. This provided Henry with the financial firepower to annex the city of Boulogne from the French in 1544, albeit at the price of England experiencing one of its greatest inflations ever.

Zoom forward five hundred years and Rishi Sunak, the Prime Minister of the UK, has ignited a controversy by referring to inflation as a tax, and further suggesting that the "best tax cut I can deliver for the British people is to halve inflation." His BBC interviewer disputed the claim, saying that inflation isn't a tax, a stance that the BBC upholds on its fact checking page.

If you recall, my previous article showed how Henry VIII's debasement functioned very much like a tax, say a new customs duty on wine or a beard tax. It did so by incentivizing people to flock to English mints to have their precious metals turned into coinage, Henry extracting a small fee on each coin. But the 21st century monetary system is very different from that of the middle ages. Is Rishi Sunak right to characterize inflation as a tax?

First, we need to better define our terms.

What do the BBC interviewer and Sunak mean by inflation? In the western world, prices have been rising at a regular pace of 2-3% each year for decades as result of central bank policy, which targets a low and steady inflation rate. Is this the definition they are using? Alternatively, Sunak and his interviewer may be referring to inflation as a *change in the change* in price. Since 2022 or so, that 2-3% rate has leapt to 8-9% all over the western world. Is it this jump that Sunak and his interviewer are talking about?

For the sake of this article, we'll assume that the conversation between Sunak and the BBC refers to the latter, a spike in the rate of inflation.

Secondly, what is meant by the word tax? Sometimes when we say that something is a tax we mean that it causes suffering. That is, inflation is taxing: it makes people's lives harder by increasing the cost of living, with salaries failing to keep up. It creates unfair changes in winners and losers.

Fair enough. But the more precise view I want to broach in this article is that inflation is actually a tax, where we define a tax as a formal charge or levy, set by the political process, that leads to cash flowing from the population to the government.

What does the data show?

Interestingly, a surprise jump in inflation leads to the very same effects as a new tax. All things staying the same, a new tax leads to an increase in government revenues. This improves the government's fiscal balance, or the difference between its revenues and expenses. A recent IMF paper by Daniel Garcia-Macia using data from 1962 to 2019 shows how an inflation shock typically achieves this exact same end result, boosting government revenues and improving its fiscal balance. This effect lasts for a few quarters, even up to two or three years, then recedes.

The IMF's chart below breaks down exactly how an inflation shock tends to improve government finances using quarterly data going back to 1999:

Charts source: IMF

Total tax revenue (the first panel) immediately begins to rise after the inflation shock at about the same rate as inflation.That's because most taxes are set by reference to values or prices, say like the prices of goods and services, or the price of labor, or the value of corporate profits. Since inflation pushes these amounts higher, this gets quickly reflected in tax revenues.

Income taxes and profits taxes (the second panel) rise particularly fast. Inflation is presumably pushing tax payers into higher income tax brackets, a process known as "bracket crreep," and so the government very quickly starts to collect a proportionally-larger amount of income tax.

Meanwhile, the government's total expenditures, the third panel, typically stay flat or only marginally rises in the quarters after the inflation shock hits. Notably, the amount of wages that are paid to government employees and social benefits (panels 4 & 8) tend to fall.

The net effect is an improvement in the government's fiscal balance. More specifically, for a 1% increase in inflation, the government's overall balance tends to improve by about 0.5% of GDP. And so an inflationary shock ends up at the same endpoint as a new tax: higher revenues and a better budget. That doesn't necessarily mean that inflation is itself a tax. Taxes have a degree of intentionality. They get implemented through a political process that has a certain set of goals in mind. By contrast, the extra revenue that an inflation shock raises is often (though not always) accidental, the result of external forces rather than political decision making.

So while it may not fall under the dictionary definition of a tax, the tax implications of a modern inflation shock resemble that of a new tax.

Everything I've written above applies to an inflation shock, say a rise from a 2-3% to 8-9%. Next I want to show that even constant 2-3% inflation can have the same revenue implication as a tax. Here's how.

Banknotes and seigniorage

Governments usually have a monopoly over the issuance of two key financial instruments: banknotes and settlement balances (also known as reserves). We all know what banknotes are, but what are settlement balances? Commercial banks find it useful to keep a stock of settlement balances on hand to make crucial large-value payments to other banks. The central bank, which the government controls, is the monopoly provider of these balances. (Sometimes banks are required by law to keep a a fixed number of settlement balances on hand, often above and beyond their day-to-day needs, a policy referred to as required reserves.)

Historically, interest rate on both types of central bank-issued money have been set at 0%. At the same time, the rates on short-term credit instruments (Treasury bills, commercial paper, bankers acceptances, etc) are determined by the market, typically hovering at a positive rate ranging between 0.25% to 5% over the last thirty years. These yields are priced to compensate investors for inflation.
The interest rate gap this gives rise to allows central banks to earn a steady stream of revenues, borrowing at an artificially cheap rate of 0% from both the banknote-using public and banks, and reinvesting at, say, 3%. Most of the revenues that the central bank collects from this interest margin flows back to the government. Economists usually refer to these revenue stream as seigniorage.

So seigniorage performs the same function as a consumption tax or an income tax: it takes resources from the public and gives it to the state. Likewise, a reduction in seigniorage would be very much like a tax cut.

If politicians wanted to, they could do away entirely with this form of raising government revenues. They have two ways of going about this. One way would be to have the central bank reduce price inflation to zero. By doing so, the interest rate on short-term credit instruments like Treasury bills would also fall to 0%, or thereabouts, since these instruments no longer need to compensate investors for inflation. And so the gap between the 0% rate at which central bank fund themselves and the rate at which they reinvest would cease to exist, seigniorage effectively shrinking to zero.

Over the last few decades, governments have taken a second route to removing seigniorage: they have begun to pay a market-linked yield on settlement balances. Canada, for instance, adopted this policy in 1999, and the Bank of England did so in 2006. By paying a market-based return, central banks no longer extract seigniorage from banks by forcing them to hold 0% assets. 

However, that still leaves banknotes as a significant source of seigniorage. We can calculate how much the UK government roughly earns from banknote seigniorage. With £95 billion in banknotes outstanding in October, and interest rates at 5.1%, the Bank of England's banknote-related seigniorage comes out to around £5 billion per year, much of which flows back to the government. That sounds like a lot, but it's only a small chunk of the £790 billion in taxes the UK government collected last year.

Banknote seigniorage isn't set in stone. It's a policy choice. If governments wanted to, they could reduce this form of seigniorage by paying interest on banknotes. One way to go about this would be to introduce a banknote serial number lottery. This lottery would offer around £5 billion in cash prizes to holders of winning banknote serial numbers, equating to a 5% interest rate on banknotes. Doing so would be akin to enacting a tax cut on British citizens.

To sum up, the fact that both an inflation shock and steady 2-3% inflation have implications for government revenues suggests that while inflation may not quite qualify as a tax, it is certainly tax-like.