Friday, April 19, 2013

A rush for US paper dollars: the rejuvenation of the world's most popular brand

Here are Paul Krugman and James Hamilton on the renewed demand for dollar bills.

So what's behind the soaring demand for US paper dollars? A simple strategy for getting a grasp on US data is to compare it to the equivalent in Canada. Comparisons between Canada and the US serve as ideal natural experiments since both of us have similar customs and geographies. By controlling for a whole range of possible factors we can tease out the defining ones.

The chart below shows the demand for Canadian paper dollars and US paper dollars over time. To make visual comparison easier, I've normalized the two series so we start at 10 in 1984. On top of each series I've overlayed an exponential trendline based on the 1984-2006 period. I've zoomed in on 1997 for no other reason than to provide a higher resolution image of the typical shape of cash demand over a year.

Some interesting observations:

1. Not a huge surprise, but the demand for US paper has been accelerating far faster than the demand for Canadian paper. As James Hamilton points out, this is no doubt due to the huge transactional demand for US dollars overseas. The emerging countries in which US paper is demanded often have high growth rates, and their requirement for transaction media is correspondingly elevated. Unlike greenbacks, Canadian loonies are only demanded in Canada. As a slow-growth country, we don't require rapidly expanding amounts of physical transactions media.

2. Zooming in on any given year (I've chosen 1997) the demand for Canadian paper is far more jagged than the demand for US paper. Why is this? My guess is that the demand for US paper is diffused across multiple nations with diverging business practices and cultures. The demand for Canadian paper, on the other hand, is tightly linked to specific Canadian customs, holiday seasons, and payroll scheduling practices. The overseas demand that smooths out and counterbalance the peculiarities of domestic US paper demand don't exist for loonies.

3. There are some neat patterns in the chart. No, not all cash is demanded by criminals. There's always a cash spike at Christmas/New Years, and if you look carefully you can see jumps in Canadian cash demand coincide with major holidays, including Thanksgiving and the September long weekend. As Lenin once said, give me data on your nation's money supply and I can tell you when its holidays are. And note the huge Y2K-inspired rush to hold paper. Cash is still the ultimate medium for coping with raw uncertainty.

4. US paper demand started to slacken relative to trend in the early 2000s. One might be tempted to blame technological advances or changes in US preferences over payment media for slowing demand. Cash is a dinosaur, right? But this can't be the case. Canadians benefit from the same technologies as the US, nor do payment preferences change when one moves from 50 miles south of the 49th parallel to 50 kilometres north of it. If technology or preferences had changes, then Canadian cash demand would have deteriorated too, but as the chart shows, it continued to rise on trend. The best explanation for the US dollar's divergence from its long term growth just as Canada hewed to its trend is that foreign demand for US paper began to decline.

It's a reasonable explanation. Around 2002, the value of the US dollar begin a long and steady deterioration against most of the world's currencies, in particular the euro. It's very probable that consumers of the US$ brand punished the brand owner, the Federal Reserve, by returning dollars enmasse to their source, thus reducing the supply of paper dollars (or at least reducing its rate of increase). As incontrovertible proof, I submit exhibit A—a 2007 video of Jay-Z flashing euros instead of dollars (skip to 0:51).

5. So it seems to me that from 2003-2008 there was a mini run on the Fed by overseas cash holders. What Jaz-Z doesn't show is the process by which US dollars would have refluxed back to the US. Euroization, or de-dollarization, goes like this. A foreigner goes to their local bank to trade US dollars for euros. The local bank, flush with dollars, puts this paper on a plane for redeposit at their US correspondent bank in New York. The New York bank, which now has too much vault cash, loads these dollars into a Brinks truck and sends them to the New York Fed. And the FRBNY shreds the notes up.

This mini run would have put downward pressure on the federal funds rate. Here's how. Having accumulated excess cash from overseas, US banks would have sent this cash to the Fed in return for reserves. But now these banks have excess reserves. Desperate to get rid of them, they all try to lend their reserves at once, driving the federal funds rate down. To ensure that the federal funds rate doesn't fall below target, the Fed would has to sell treasuries in order to suck in reserves, thereby reducing the oversupply in the federal funds market and keeping the fed funds fixed.

The lesson being, when folks like Jambo in Zimbabwe and Julio in Panama get distraught about the quality of their Ben Franklins, the effects of their unhappiness will be felt, with some delay, all the way back at the Fed's open market desk.

6. US paper demand has since rebounded. Paul Krugman posts a chart that shows a massive accumulation of US cash holdings relative to GDP beginning in 2008. But Krugman's chart overstates the effect by constricting his time frame. As my chart shows, the rate of growth in US paper has only returned to the trend it demonstrated in previous decades.

Krugman attributes this increase in dollar holdings to the fact that the US is in a liquidity trap. When rates are near zero, people have no problems holding zero-yielding cash. I'm not so sure about his explanation. Canada had incredibly low rates for a few years, yet as our chart shows, Canadian paper never budged from its trendline growth. The same goes for the Euro. Rates have been low there, but we haven't seen a flight into paper money. Because cash is inconvenient and bulky, rates have to go pretty far below zero before people flee to paper.

No, the more likely explanation for the rebound in the US paper outstanding is the rejuvenation of the US dollar brand. The ECB has had to deal with waves of negative publicity for the last few years. Given the alternatives, the world wants to hold Benjamins again. This seems to be borne out in the chart below, which shows the ratio of ECB-to-Fed banknotes in circulation.

The US dollar, it would seem, is back. Cash holdings are only one sign off a currency's hegemony. It would be telling if there's also been a rebound in the use of the dollar to denominate bonds and other debt instruments, as well as increased holdings of US dollar-denominated assets in the reserves of major central banks. The US's "exorbitant privilege", as Barry Eichengreen calls it, continues apace.

Note: As I was writing this, I stumbled on a paper by Ruth Judson via James Hamilton called Crisis and Calm: Demand for U.S. Currency at Home and Abroad from the Fall of the Berlin Wall to 2011. And what do you know. She uses Canada as a foil for determining US cash demand, just like I did. I haven't read it yet, but am quite looking forward to doing so and am willing to yack about it in the comments.

On Lenin, read White & Schuler.


  1. Lovely post.

    I laughed at the Lenin "quote". (A little parody of Keynes?)

    1. Yep. If a big shot like Keynes can get away with coining fake Lenin quotes, we should all be allowed ;)

  2. I have a somewhat unrelated question if you don't mind. Do you agree with the backing theory of money? And if paper money is backed, what is the right way to think of its liquidity premium?

    1. See some of the discussions Mike Sproul and I have had in the comments here, here, here, and elsewhere.

    2. I'll see if I can field that one. JP and I are in about 90% agreement about the backing theory. JP is more agnostic about it, while I am more inclined to say that the forces governing the price of dollar bills are no different from the forces governing the price of bonds.

      As far as the liquidity premium, JP once said that the liquidity premium is a small slice on top of fundamental value. I'm OK with that, but of course competition from rival moneys would make that premium small, or zero. I doubt that JP would go so far as to set the liquidity premium at zero. He did, after all, call his blog "moneyness", which implies that he thinks that moneyness matters to the value of money.

    3. That was very helpful, thanks! It seems like the answers to my questions are 1, yes and 2, very small. I just don't get the right way to think about this. How can something that is extremely liquid, like paper money, have a very small liquidity premium?

    4. Think of a paper IOU (a ‘dollar’) that promises 1 oz of silver, payable in 1 year but inconvertible until then. If the interest rate is 5%, and printing and handling costs of that dollar are zero, then that dollar will sell for .95 oz today and will rise to 1 oz at year-end. The issuer will earn zero economic profit on that dollar.

      Next, think of a convertible paper dollar that promises 1 oz of silver, payable anytime. If the interest rate is 5%, and printing and handling costs of that dollar are zero, then the issuer can sell the IOU for 1 oz at the start of the year, lend the 1 oz for 1 year at 5%, get repaid 1.05 oz at year-end, buy back the dollar for 1 oz, and earn a profit of .05 oz. This profit will attract a rival money issuer who could, for example, offer a rival dollar that says “IOU .99 oz. +1% interest, payable anytime”. That dollar will start the year worth .99 oz. and rise to 1 oz at year-end. The issuer could sell the dollar for .99 oz today, lend the .99 oz for 1 year at 5%, get repaid 1.04 oz in 1 year, buy back the dollar for 1 oz, and earn a profit of .04 oz. This profit will attract another rival who will offer dollars that say “IOU .98 oz. +2%”. That rival can earn a profit of .03 oz, and the rivalry will continue until someone issues dollars that say “IOU .95 oz +5%”. Now the profit to the issuer will be zero, so we are at a stable equilibrium. Note that in zero-profit equilibrium, the convertible dollar is worth the same as the inconvertible dollar throughout the year.

      Now we get to the liquidity premium that you mentioned. Suppose that a dollar (either kind) sold for .96 oz today, instead of the .95 oz. dictated by the 5% interest rate. A money issuer could earn an arbitrage profit of .01 oz. by issuing and selling a dollar today for .96 oz, lending the .96 oz at 5%, getting back 1.01 oz at year end, and buying back the dollar for 1 oz. That’s why I think the liquidity premium must be close to zero.

      What if the dollar’s printing and handling costs were .03 oz/year? In that case the dollar won’t yield 5%, but only 2%. This means the dollar that promises 1 oz at year-end will sell for .98 oz today, and will rise to 1 oz by year-end. (Any other price creates arbitrage opportunities.) People are willing to tolerate the dollar’s 2% return (as opposed to regular bonds’ 5% return) because the dollar has a convenience yield of 3% or more. The price of .98 makes it look like the dollar has a liquidity premium of .03 (=.98-.95), and in a sense that’s true. But note that if the printing and handling costs fell to zero, the dollar would fall to .95 oz and the liquidity premium would disappear. It turns out that the size of the so-called liquidity premium is entirely dictated by printing and handling costs.

      Keep in mind that I’m using a 1-year time horizon. The premium would be much bigger for a long time horizon.

    5. Thanks a lot! That pinpoints where it comes from. If arbitrage opportunities keeps it this small that must mean that wherever liquidity premiums persists, for any financial instrument and above the costs of issuing, it cannot be arbitraged away, or for some reason this takes a lot of time?

      Thanks again for taking your time to answer. I've read your site about real bills/backing theory and I found it both easy to follow and very intuitive. Would you say the real bills view is gaining support?

    6. With few exceptions, I am forced to agree with Charles Calomiris’ (1997) assessment that “to my knowledge, the real bills doctrine has no current advocates.”

      Those "few exceptions" would include people who have written in favor of the "fiscal theory of the price level" (e.g., John Cochrane, Thomas Sargent, etc), as well as people I've come across on the internet who are at least open to the idea of the RBD (JP, Michael Rozeff, Jacques Raiman, David Perrins, David Glasner, and maybe about a dozen others). The people I've argued with the most are Nick Rowe and Scott Sumner, both of whom reject the real bills view.

    7. Ok, I know the new austrian school of economics is also in support of RBD.

    8. The size of liquidity premia seems to me like an empirical question. I'm interested in thinking up ways to measure liquidity premia and protect oneself against liquidity risk. Like here.

    9. The stock market idea sounds interesting but how can price appreciation be part of a stocks real return and be filtered out in one product, and liquidity premium in another one? If a stock is made non tradable I suppose that would filter out all value that isn't the value of the expected income stream at the time the stock was purchased.

  3. Another nice post.

    Maybe some of the difference between the US and Canada can be explained by the fast growing grey economy (which is illegal, not "criminal" activity) in the US?
    Since the number of employed Americans is constantly decreasing, it may be that the rest are increasingly participating in the cash-based grey (and of course black) economy.

    A problem with the euro is that even citizens of countries that aren't failing have to lower their euro holdings in order to insulate themselves from their dear leadership and F-PIIGSS.

    Eurocrats are working on introducing full scale financial repression and the use of cash is not only discouraged, but criminalized (see the introduction of maximum amount allowed for cash-based transactions under the euro-marionette Monti, etc.)
    Germany (of course - who else) is actively lobbying for the cancellation of EUR500 notes.

    No wonder people are getting rid of their euro banknotes.
    We know that after the incident in Cyprus billions were printed (I blogged about that) to pacify nervous depositors, but if they're any smart they'll be getting rid of those real fast (and convert them to anything else but euros).

    1. Your first point is an interesting explanation, and it's testable. As we move out of the recession we'd expect to see a decrease in US cash outstanding. Mind you, Canada's unemployment spiked in 2008-10 yet there was no visible change in the pattern of cash outstanding, so I'm not sure strong the 'grey market' effect is.

  4. The evolution of god: