OPINION: The SARB’s understanding of costs of QE critically flawed
JOHANNESBURG – On June 18, the Governor of the SA Reserve Bank (SARB), Lesetja Kganyago, delivered a lecture at the Wits School of Governance titled, “The South African Reserve Bank, the coronavirus shock, and ‘the age of magic money’”.
In that lecture he sought to provide “more clarity around the mechanics of asset purchases, quantitative easing (QE), and the ‘zero lower-bound’”. In the process, the Governor made a number of statements which require critical examination.
In this piece, I focus on his incorrect claims that those who call for QE “take the mistaken view that QE is ‘free money’”.
I also show that the SARB holds a critically flawed understanding of how QE is financed. In fact the SARB’s flawed understanding implies that QE is “free money”. Hence the erroneous claim that QE will bankrupt the SARB must be rejected. To simplify my argument, I will treat cash and reserves as if they are one and the same thing.
QE is the process by which a central bank increases liquidity in the economy by changing quantities of assets and liabilities.
One way to do this is to buy securities, such as government bonds, say from commercial banks. After selling their government bonds to the central bank, banks will have excess cash which they could lend cheaply to their customers. This excess cash causes interest rates to fall as banks compete to get borrowers because holding excess cash does not earn any nominal return for banks. The Governor argues that this could cause inflation, thereby violating his mandate of maintaining inflation on target.
To prevent inflation, the central bank will have to divert the cash away from being lent by banks. The central bank then opens interest-bearing accounts for banks to deposit their excess cash. Therefore, as the Governor says, the SARB can prevent the excess cash from causing inflation by “paying banks to deposit their reserves back with us”. Thus, upon getting the cash from the sale of their government bonds, banks would keep the excess cash in their accounts at the central bank and earn interest.
The Governor points out that these interest costs on the SARB would be large.
He says, “if we as the SARB bought R500 billion government bonds, at par, and then sterilised them at the repo rate, we would be insolvent in about a year”.
To sterilise is to prevent the R500bn excess cash from being lent out by banks by opening interest-bearing accounts for banks at the SARB, as already explained. The Governor estimates that these interest costs will be R19bn per year. The SARB’s capital plus accumulated reserves are R20bn. The Governor erroneously concludes that QE would bankrupt the SARB, “in about a year”.
Kganyago erroneously argues that QE is appropriate only on one condition; when inflation is threatening to be negative so that the repo rate is pulled down to zero.
According to the Governor, this situation will not force the SARB to pay interest on the excess cash because interest rates would be zero anyway.
Secondly, the excess cash would not pose an inflation threat because inflation would be far below target. Since these conditions do not hold, QE is not appropriate. My argument is that this expresses a defective understanding of how QE is financed and it reveals some logical incoherence.
Firstly, by requiring interest rates to be zero in order to implement QE, the SARB holds the “free money” view it claims to oppose, i.e. that QE is money issued at 0 percent interest rate by the central bank.
By requiring the inflation rate to be near zero, the SARB also fails to see that for an emerging market such as South Africa at that point the unemployment rate would be at even more catastrophic levels. The reason for this is that there should be prolonged and excessive demand shortage to depress inflation way below target. This is equivalent to increasing unemployment because firms would not be able to sell what they produce.
Secondly, the claim that my colleagues and I have taken a mistaken view that QE is “free money” is baseless.
In footnote 19 of his speech, Kganyago cites me when I said the correct interest rate at which QE in an emerging market such as South Africa should be implemented is the one that equals the rate ruling in advanced economies plus a sovereign risk premium and this is, in general, not zero.
As I have argued repeatedly elsewhere, this rate protects the external value of the currency by guaranteeing a stable exchange rate.
However, the Governor appears to have not understood the implications of this basic proposition for the design of QE.
In my previous pieces, I took it for granted that the SARB officials would understand that this basic proposition implies that it is this rate that should be used to pay interest on the excess cash. This is not to prevent commercial banks from lending to the public, but it is to prevent banks from using this excess cash to acquire foreign currency assets, thereby leading to currency depreciation. The QE that I propose comes at a cost. It is, therefore, not correct for the Governor to say QE proponents in South Africa have not appreciated the costs of QE.
Thirdly, the SARB’s monetary policy mandate is to protect the value of the currency in the interest of balanced economic growth.
When there are prolonged disinflationary forces at play as there are currently, such as the deep and persistent under-utilization of capacity, rising unemployment rate and a fall in the prices of inputs such as oil, inflation gets pushed below target.
What then happens is that a possibility arises where the policy rate drifts below the one that is consistent with a stable exchange rate. What the SARB needs to do under this situation is to set the repo rate at the level that is consistent with the stability of the exchange rate, and then embark on QE. Will this bankrupt the SARB, as Kganyago says? No, and this is why.
Suppose, as the Governor assumes, the SARB purchases all of the R500bn government bonds held by the banks. This will add R500bn excess cash on the banks’ accounts at the SARB. The SARB will then pay banks to hold these excess reserves an interest rate that equals the sovereign risk premium plus the advanced economy rate, which is then the prevailing repo rate. At the repo rate say of 3.75 percent, this amounts to R18.75bn per annum. Kganyago leaves the story here and proceeds to conjure up a spectre of the SARB being bankrupted by QE. However, by leaving the story here a critical flaw appears in the Governor’s reasoning about the costs of QE.
The story, taken to its logical conclusion, goes like this. The government would now owe the SARB, since the bonds are now on the SARB’s balance sheet as assets. The coupon rate on those bonds, i.e. the interest rate that government pays, is about 10 percent. The spread between the coupon rate and the repo rate of 3.75 percent, which is 6.25 percent, would be profit to the SARB. This profit would amount to R31.25bn a year. The QE would therefore be self-financing by increasing the profits of the SARB and not erode them. Far from weakening the SARB’s balance sheet, QE would strengthen it by increasing the SARB’s accumulated reserve fund, while achieving monetary policy and financial stability goals.
What is the effect of this QE on the government? Section 24 of the SARB Act says the SARB should remit 90 percent of its surplus to the government, after making some provisions. Assume these provisions plus an allocation to the SARB’s reserve fund take up 0.25 percentage points of the spread.
The SARB would then remit the 6 percent spread to the government. Effectively government would be paying 4 percent interest on its outstanding debt that is now sitting on the balance sheet of the SARB, instead of 10 percent. This would amount to an annual injection of R30bn into government coffers, thereby, contrary to Kganyago's assertion, lifting the government’s budget constraint.
I must add that the SARB has options to further assist government in paying the 4 percent interest rate on excess cash. These options notwithstanding, a QE programme designed along the lines I have outlined above would significantly ease the fiscal constraint on government and support liquidity provision to the banking sector thereby ensuring financial stability. With appropriate amendments to the SARB Act, it can be extended to ease credit conditions for the household and commercial sectors, thereby restoring confidence in the system.
Christopher Malikane is Associate Professor in the School of Economics and Finance at Wits University.