The Reserve Bank. Picture: SARB
The Reserve Bank. Picture: SARB

Unlike the three oracles who correctly forewarned Macbeth about impending doom, former SA Reserve Bank economist Vukani Mamba’s response to Prof Chris Malikane’s defence of quantitative easing (QE) is premised on a misunderstanding of monetary policy proposals (QE proposal could yield untold financial instability for SA’s economy, July 20).

Mamba claims that Malikane has proposed a coupling of the repo rate and the exchange rate, and that this is ill-conceived; that a fully sterilised QE programme is self-defeating; that excess Reserve Bank reserves in the hands of commercial banks will result in reckless lending; and that the carry trade from developed countries off the back of their own QE programmes, will wreak havoc with the SA economy if we also implement our own QE programme.

Mamba's entry point into the debate is what he sees as an impractical proposal by Malikane, who he says has, in effect, proposed to have the repo rate coupled with the exchange rate. He then bemoans the difficulty this would present for the technical model calibration process, since you would have to factor exchange rate volatility into the repo rate modelling. However, no such coupling claim, nor proposal, was made by Malikane. What he said was if the Bank desires a stable exchange rate, there is a repo rate floor below which the rate cannot be set.

Put simply, Malikane is not proposing a repo rate that dances up and down matching the exchange rate, but is stating that much as the repo rate may dance to its own tune, it cannot dance below a certain threshold. This lower threshold is driven by the idiosyncratic risk foreign investors see when they want to invest in SA. Such a risk is compactly called the sovereign credit risk.

A quantifiable approximation of this sovereign risk is the term structure of CDS (credit default swap) spreads. CDSs are liquidly traded in global financial markets. In the case where rand denominated CDS spreads are not liquidly traded, there are dollar CDS spreads that can be quantoed into rand spreads. Alternatively, bond implied CDS spreads could also be used to calibrate such a threshold.

If one traces the above exposition it is obvious that the exchange rate is not a primary driver of this repo rate threshold, nor is this threshold set with the aim of targeting some predetermined exchange rate level. All subsequent analysis Mamba makes based on this flawed understanding of the proposal cripples his argument.

At risk of putting too fine a point on this (no doubt because this repo rate floor point has been misunderstood by no less an authority than the Reserve Bank), I will hazard a metaphor. When a corporate employee is sent to a foreign country where there is life-and-limb-risk to their life, that employee receives danger pay to compensate for the idiosyncratic risk being in that particular country poses. Similarly, the salary foreign debt holders earn in SA is the interest rate on their lending, and that rate has danger pay baked into it.

Thus, if you wish foreign “employees” to remain invested in your risky country the interest rate term structure (as indexed by the repo rate) cannot collapse to zero because that would imply that the baked-in danger-pay also collapses to zero. In other words, to the extent that the repo rate is a policy guidance rate that informs the Bank's open market operations with the goal of modulating market interest rates, those market rates cannot collapse to zero.

If rates were to collapse below the danger pay threshold, much like disenchanted employees in a foreign war zone, investors would pull out because they would not be meeting their risk adjusted returns. This pulling out would destabilise the exchange rate, hence the floor threshold (implicit or explicit) acts to support the exchange rate. Malikane is not suggesting that the amount of danger pay should dance to the tune of the exchange rate, merely that the danger vpay is not zero and reflects the sovereign risk borne by foreign investors.

Mamba also argues that a fully sterilised QE, as proposed by Malikane, is self-defeating. He then entertains various degrees of nonsterilised QE because he imagines this is the only worthy possibility.

Again, Mamba is painfully adrift. The objective of QE is to shape the yield curve, specifically to lower the yields at which government is borrowing money. The objective of sterilisation is to drain the excess liquidity that has been injected into the financial system.

Draining excess liquidity need not nullify the yield curve shaping of the QE. In other words, you can fully sterilise QE funds yet still pass through lower borrowing costs to government. The difficulty of how this is possible defeated Reserve Bank governor Lesetja Kganyago in his Wits University lecture, in which he posited that draining excess liquidity by using market instruments (this process, the governor then argued, would nullify the yield curve shaping — he is correct).

The intervention Malikane made is by articulating that you can drain the liquidity through a non-market instrument (the reserve system). The advantage of this is that the governor can then choose the sterilisation interest rate (interest on reserves), which then makes the whole exercise possible, not self-defeating.

Related to this is the issue of using QE to stimulate the economy. It has been conceded by many progressives that QE alone is not enough. For additional economic stimulus we would also need to use the Bank's balance sheet. I shall again hazard a metaphor to clarify the above point. When you are mopping the floor you throw water on the floor (inject QE liquidity) then wipe it off (drain liquidity).

To a noncleaner this process seems self-defeating because you are reversing what you initially did. However, to the keen observer the benefit of such a process is reduced dirt on the floor (lowering government borrowing costs). It is trite that you would not water a garden in this way (stimulate growth), so why would you need a different watering strategy for mopping (using the Bank's balance sheet beyond the proposed QE).

Mamba also suggests that excess reserves from nonsterilised QE are going to lead to reckless bank lending. Recall from the exposition of the second claim that it was shown that Mamba entertains nonsterilised QE because he feels fully sterilised QE is self-defeating. The premise of his nonsterilised QE is flawed (as argued above) but let’s ignore this flaw and proceed to dismiss the reckless lending argument within the cocoon of its existence.

There is currently a R200bn treasury loan-guarantee scheme where the Bank advances reserves at the repo rate plus a spread. These reserves provide support for banks to onlend to Covid-19-afflicted businesses. It has been widely reported in the media — and finance minister Tito Mboweni also conceded this in his supplementary budget speech — that the uptake of this scheme has been poor. A core reason for this is that the lending conditions banks impose on businesses are too onerous (not reckless).

What this demonstrates is that there are excess reserves advanced by the Bank, and simultaneously stringent lending standards applied by banks. But Mamba says this should and would be the opposite: that banks should be on-lending these reserves recklessly.

Why is he wrong? The lending decisions of banks are adjudicated by credit committees and credit policies, which pivot on the creditworthiness of bank clients. During a recession these lending policies tend to become much more stringent because banks self-assess that from a macro perspective there would be reduced loan repayment capacity and thus more nonperforming loans. Banks then offset this reduced macro repayment capacity through provisioning profits to cover such loan losses, and also through pruning their loan books of riskier lending. This is what is meant by the phrase “bank lending is pro-cyclical”.

Separate from credit committees, the management of excess reserves and the process of haggling them in the overnight interbank market is tasked to a separate part of a bank, called the liquidity & funding team. Yes, a deluge of excess reserves would collapse the overnight rate as a result of harried haggling in the interbank market because of activity between the liquidity teams of the different commercial banks in the country. This collapse is precisely why sterilisation is necessary to support the overnight rate.

Notwithstanding this, the liquidity team does not sit in on individual client lending decisions save for normal intra-bank communication between teams about optimising that bank’s balance sheet. If the situation described above were not the case, the reckless lending of which Mamba forewarns would have appeared already with the treasury loan-guarantee scheme.

Mamba warns of how carry trade off QE in developed countries is going to overwhelm SA's financial markets, and thus any additional QE of our own would make a bad situation dangerous. This claim is made off the fallacy already explained above, that our QE would be nonsterilised. Malikane has explained that ours will be sterilised, thus draining any excess liquidity that could possibly combine with carry trade inflows.

Further, this carry trade claim is made off a flawed understanding of the repo rate floor that is proposed. Mamba says the repo rate proposal would attract a deluge of carry trade and would thus possibly be destabilising. However, as already explained the proposal is for a repo rate floor and not for a repo rate that tracks the exchange rate with the aim of achieving a target exchange rate.

This is important because the aim of the repo rate floor is not to maintain a funded current account but rather is a one tool in a suite of mechanisms that can be used to wean SA off its dependence on the carry trade (by progressively lowering the repo rate in a manner that attenuates foreign inflows into the local economy). Currently, somewhere in the order of 40% of rand denominated government bonds are owned by foreigners.

Thus, to the extent that the repo policy rate informs Bank open market operations used to modulate the market-rates term-structure, the repo rate floor that is proposed is not aiming to increase foreign inflows but rather prevent them from collapsing precipitously in the short-term due to the disappearance of “danger-pay” (thus destabilising the exchange rate).

In fact, (and a bit off topic) there is a present and real problem with activity in the forex spot and forex forward markets, which frustrates the Bank’s repo rate guidance policy. This problem is the excess liquidity with which foreign inflows saddle the overnight market. This means that much as the Bank may regulate domestic credit conditions, foreign inflows arrive like an unexpected guest and drown the interbank market with liquidity.

The reason this problem is pertinent to the proposal by Malikane is because a nonzero interest rate on reserves policy by the Reserve Bank could actually help it modulate interbank lending even in the face of unexpected and inconvenient foreign guests. This is what makes the policy proposal to use nonzero interest on reserves so versatile: it could help the Bank solve other problems it currently has with regulating the rates in the interbank market.

There are quantitative limits to the exact amount of QE the Bank can undertake. These derive from both the statutory quotas governing the Bank and, importantly, from the operational structure of the commercial banking system. However, such limits are not Mamba's concern. He sounds like a patriot for the concerns he has raised about SA and the stability of its institutions (of which the banking system is one). His concerns are well received, but as patriots we are enjoined to “claim no easy victories”.

• Magolego prices exotic and bespoke derivatives at a local bank. He writes in his personal capacity.

Would you like to comment on this article or view other readers' comments?
Register (it’s quick and free) or sign in now.

Speech Bubbles

Please read our Comment Policy before commenting.