Bank Of Canada: Repo Printer Go Brrr!

The Bank of Canada has completely restructured its balance sheet in response to pandemic stresses, causing it to resemble that of the Federal Reserve (above figure). The size of the balance sheet has exploded, creating yet another time series terminated with a vertical line. If we look at the assets, we see that repurchase agreements ("repos") went from a nearly insignificant asset to about half the balance sheet (as happened in the Financial Crisis of 2008).

(Note: Most of this article covers background material, for readers who are somewhat interested in monetary operations. The concluding section has the short version of what the Bank of Canada is doing. The briefest possible summary is that the Bank is just operating to keep the bond markets functional, so as to avoid a seizure of wholesale finance.)

Another set of notable developments are happening on the liability side of the balance sheet (above). If we looked at the pre-2008 Bank of Canada balance sheet (top panel), almost all of the liabilities were accounted for by currency notes in circulation (dollar bills in denominations of $5 and greater, currently composed of plastic). After the Financial Crisis, the Government of Canada has kept a large balance at the Bank of Canada for prudential[1] reasons. If we add those two liability items together, they still nearly represented almost the entire side of the balance sheet -- until 2020.

The bottom panel shows where the liability growth is: formally known as "Canadian dollar deposits, members of Payments Canada," but I have labelled this as "Banks" (under the argument that non-banks are not significant members of Payments Canada). Canada abolished bank reserve requirements in the 1990s, and instead banks were expected to have a $0 balance with respect to the Bank of Canada/payments system at the end of the working day. This was a key part of the job description for bank treasury departments: hit that zero balance at the end of day, given that there are massive gross flows in and out of the bank throughout the day. The Bank of Canada posts official deposit/lending rates on positive/negative bank balances, but the expectation was that if private banks missed the zero target on a sustained basis, inspectors would come around asking very awkward questions to bank management. That is, although those posted rates acted as a corridor for private sector inter-bank rates, the expectation was that there should be almost no transactions.

Since required reserves are zero, hitting that target at the aggregate level is very easy for the Bank of Canada: they just have to do the opposite of Government of Canada ("Treasury") transfers, as well as the flow of currency notes from private banks. After that, private banks with deficits just need to get their hands on the surpluses at other banks in the inter-bank market (by any means necessary).

What has happened in 2020 is that banks have parked deposits at the Bank of Canada (which happened to a smaller extent in the Financial Crisis, as seen in the chart).

Simplified Framework of Government Finance

In my book Understanding Government Finance (link to online bookstores), I discussed what I termed the "Simplified Framework of Government Finance" (Section 3.3). I used this framework to explain monetary operations, which feature prominently in Modern Monetary Theory (MMT). The Simplified Framework was based on how the Bank of Canada operated in practice, which deviated from how the American Federal Reserve operated.

Historically, the main difference between the Bank of Canada and the U.S. Federal Reserve was bank reserve requirements. Since American economics 101 textbook authors insisted on indoctrinating students with money multiplier stories[2], the Federal Reserve clung to these archaic rules, which requires banks to hit a particular level of settlement balances at the end of reserve periods. This makes monetary operations more complex, which in turn infected most MMT primers. (It is unclear whether the recent abolition of reserve requirements in the United States will show up in mainstream economics 101 textbooks.) I got rid of reserve requirements in my Simplified Framework, resulting in, umm, a simplified framework in which to study monetary operations.

Repo!

The next qualitative difference that used to exist was the use of repurchase agreements, commonly termed repos (since that sounds cooler). If one looks at the top panel of the first figure in this article, one sees that the size of the Bank of Canada's balance sheet used to rise slowly over time (with a small hiccup around the Financial Crisis). As the top panel of the next figure showed, that balance sheet size was almost exactly equal to the amount of currency in circulation (and then, the Government of Canada's balance in the Receiver General's fund). 

Currency demand is relatively stable over time (with seasonal effects); supporting low denomination transctions by legitimate businesses, and larger denomination transactions in the underground economy (as in the "legitimate businessman's club"). (United States currency is used outside the United States as a store of value -- explaining why the $100 bill is issued in large amounts -- whereas the Canadian dollar has less cachet on that front.) This demand stability meant that the Bank of Canada's balance sheet size was stable, and there would be little need to shrink its balance sheet in response to cyclical developments. This allowed the Bank of Canada to steadily grow its balance sheet by the straightforward tactic of buying Government of Canada bonds outright.[3] 

Conversely, since bank deposits follow lending activity, the Federal Reserve needed to have the capacity to shrink its balance sheet. This meant outright purchases of bonds is less attractive, since the central bank does not want to deal with the accounting headache of capital losses. Instead, part of the balance sheet growth is via asset purchases that can be easily reversed without a capital loss -- which is literally what a repurchase agreement is.[4] As such, a significant portion of the Fed's balance sheet always consisted of repos, whereas the Bank of Canada saw little need for those instruments.[5]

The newfound love of repos by the Bank of Canada gives a quantitative metric to the extent that the Bank views its balance sheet growth to be temporary.

Why Do This?

Having covered the explanation of what is happening, we can now turn to the more interesting question: why is the Bank of Canada doing this?

One could slog through Bank of Canada communications, and distill the official answer. However, the explanation seems rather obvious: the collapse in economic activity caused the financial system to be looking out over the abyss, and drastic measures were needed to put a net in place.

The risks to the private sector are straightforward. Although the CMHC has generously guaranteed the entire stock of high-risk residential mortgages, there are still debts issued by non-financial firms, as well as unsecured household debt. Cascading failures here are obviously dangerous.

However, the extent of the downturn in activity is raising even scarier risks -- provincial finance. The Canadian provinces are in the forefront of the Canadian welfare state -- but they do not have their own central banks. One may note that Euro area sovereigns are often compared to Canadian provinces. Much like the ECB, the Bank of Canada has no choice but to throw out the notion of moral hazard, and make sure that the provincial bond market remains liquid.

The mechanisms are both direct and indirect: the Bank of Canada appears to be buying everything that is not nailed down, and is extending easy credit to the Banks to allow them to finance positions in governmental bonds. Meanwhile, those banks are parking their cash at the Bank of Canada, keeping their risk measures nice and controlled.

This should not be thought of as "saving the economy," at least directly. Instead, it is preventing a mindless collapse of wholesale funding markets. This means that economic distress will reflect problems in the real economy, rather than the real economy being dragged down by the dimwitry of the financial sector (as was the case in 2008). Unfortunately, there are serious problems in the real economy, as there is a surplus of expensive hydrocarbon sources, and some industries are not going to be functional in the post-virus "normal." No amount of balance sheet wizardry by the Bank of Canada can address those issues; at best, fiscal policy can be used to counter-act the effects of the needed changes to activity.

Footnotes:

[1] The Government of Canada ("the Treasury" if we map to American parlance) issues long-dated debt as a counter-part of that balance at the Bank of Canada. The government launched a new programme to lend to banks on a short-term basis, since this cash balance is far in excess of actual needs. The net result is that the Government of Canada is borrowing at a high rate in long-term debt markets in order to be able to lend overnight to banks (and the banks themselves finance positions in long-term debt). This is considered "prudential financial management" in the neoliberal era, and economists employed by banks feel that this is very much in the national interest.

[2] "Practical men who believe themselves to be quite exempt from any intellectual influence, are usually the slaves of some defunct economist." J.M. Keynes.

[3] The Bank of Canada purchases Government of Canada bonds outright during auctions. If this were more well known, many Teutonic economists would need access to fainting couches.

[4] The central bank enters into an agreement with a counter-party -- presumably a primary dealer, which are mainly bank subsidiaries in 2020 -- to buy a bond at a certain price, with a matching obligation by the counter-party to buy it back at another price at a fixed date in the future. (If the next day, it is an overnight repo, otherwise it is a term repo.) This is economically equivalent to the central bank lending to the counter-party using a bond as collateral, and the pricing convention is based on the implied rate of interest -- the "repo rate."

[5] If one looks at the data table for the Bank of Canada balance sheet, repos are broken out as an item. Previously, repos were part of "Other Assets" (and are treated as such before 2007). I had to contact the Bank of Canada when writing Understanding Government Finance to verify that repos were part of "Other Assets," and I probably would need to update that portion of the text if I do a new edition.


(c) Brian Romanchuk 2020