Key highlights:
- Credit spread trends: Over the past decade (excluding Covid years), credit spreads in the local debt capital market have tightened, despite increasing credit risk.
- Credit risk factors: South Africa's economic environment has deteriorated, with slow GDP growth, high unemployment, rising government debt, and multiple sovereign and corporate credit rating downgrades.
- Supply and demand dynamics: Limited new issuance of credit bonds, combined with increasing investor demand, has contributed to downward pressure on credit spreads.
- Public auction vs private placement: Public auctions drive competitive pricing, while private placements offer issuers quicker execution and greater control over pricing.
- Dutch auction mechanism: The local debt market uses sealed bid Dutch auctions, where all successful investors receive allocations at the final clearing price, which may not always reflect fair value.
- Bid undercutting: Investors strategically place low-volume aggressive bids to secure allocations at more favourable clearing spreads, influencing market pricing.
- Price guidance influence: Banks provide price guidance based on past auctions, but investor overreliance on this guidance may distort fair value assessments.
For much of the past decade, except the Covid years, local debt capital markets have been characterised by credit spread tightening. A credit spread is the additional interest rate above a reference rate that issuers must pay investors to buy their debt (bonds). The spread aims to capture the risk premium that must be paid to investors to compensate them for taking additional risks, for example, credit risk and liquidity risk. The underlying reference rate plus the spread will make up the total interest rate the issuer pays the investor. All else equal, issuers want this interest rate to be as low as possible (cheaper debt) and investors want this to be high relative to the underlying risk (better risk-adjusted returns). If market pricing represents fair value, issuers of poorer credit quality are expected to have a higher interest rate (higher spread over reference rate) than issuers of better credit quality, all else equal.
Hasn’t credit risk increased?
Looking back from 2020 until now, there is a strong argument to suggest that credit risk in the economy has increased. Real GDP growth has been lacklustre, unemployment remains high, and the consumer is under pressure because of elevated household debt in a high-interest rate environment, coupled with high inflation. Government debt has increased sharply from around 57% in 2019 to around 80% today. And the fiscus is increasingly under pressure to boost the economy and provide funding to ailing state-owned entities. Ratings from credit rating agencies reflect an increased credit risk environment. In March 2020, Moody’s downgraded South Africa’s sovereign rating from investment grade to sub-investment grade and issued a further downgrade in November that same year. Many corporates also suffered the same fate, with downgrades to their credit ratings. Given this backdrop, why have we seen a consistent decrease in credit spreads across most sectors of the local debt capital market? We believe the answer is not only driven by supply and demand but also by how auctions are used to influence spreads. In other words, credit spreads may not be reflective of the fair value of risk premia they aim to capture.
Demand exceeds supply
The supply of credit bonds in the market has not grown meaningfully over the past decade. In 2019 the highest volumes in the bond market, of around R174bn, were issued and then Covid dented issuance in 2020 to merely R92bn. This has slowly recovered over the past four years, with the 2024 issuance totalling R165bn. However, most of this issuance is rolling existing debt as they mature and not new, meaningful net issuance. This lack of supply is linked to the poor economic backdrop of the country. Real GDP growth in South Africa has been muted, which has meant that businesses have not grown and are investing less in capital expenditure to expand their operations. This results in businesses requiring less funding from the market. Demand for credit bonds on the other hand has been increasing. Bonds provide a stable return in times of uncertainty and high interest rates can make bonds relatively more attractive than other asset classes. We have seen bond auctions attracting more participants and auction subscription rates (volume of bids received versus volume of bonds placed) have been increasing. Expectedly, this change in supply/demand dynamics has resulted in downward pressure on credit spreads which, on its own, represents a decrease in the fair market value of risk premia. However, we believe that the actual auction process has also contributed to further spread tightening beyond these fair values.
Public auction vs private placement
Investors access credit bonds via two mechanisms, a public auction or private placement. As the name suggests, a public auction is available to all investors, who must submit bids (volume and spread) for the credit bonds on offer. The auction process is discussed further below. A private placement is a non-public, bilateral agreement between the issuer and investor(s) to buy a specific amount of bonds at a negotiated and pre-determined price. Issuers choose between coming to market via public auction or private placement for several reasons:
Auctions in the local debt capital market
The type of auction that is used in the local bond market is called a Sealed Bid Dutch Auction. A Sealed bid means an investor will submit their best bid without knowing any of the other competing bids, i.e. there is no feedback on how good or bad their bid is. A Dutch auction is a style of auction that fills up the required amount of stock (credit bonds in our example) from cheapest to most expensive (from the issuer’s perspective) until the required amount of bonds is fully allocated. Every successful investor will then get that bond at the last successful price. The interest rate that is consistent with that last successful price is known as the clearing (interest) rate. The associated spread after subtracting the reference rate, is known as the clearing spread. For example, let’s assume MTN are looking to place R100m in a bond and host a public auction where investors provide the following bids:
- Investor 1: R10m at 8.5%
- Investor 2: R50m at 9%
- Investor 3: R50m at 10%
- Investor 4: R20m at 10.5%
MTN then applies Dutch auction principles and allocates from cheapest to most expensive until they fill R100m. They will fully allocate R10m to Investor 1, R50m to Investor 2 and partially allocate R40m (= R100m - R10m - R50m) to Investor 3. Given the last successful bid was at 10% (Investor 3’s bid), all successful investors (1, 2 and 3) will all receive their allocations at 10%. Investors 1 and 2 were prepared to offer a lower interest rate but have now benefitted from the higher clearing interest rate of 10%. Investor 4 was unsuccessful and received nothing.
Interestingly, the term “Dutch auction” originates from an auction style used in the 17th century in Holland’s popular tulips market. Tulip bulbs were highly sought after and the exchange figured that the best way to sell them quickly would be with as few bids as possible, while still getting the best possible price. This style of auction is effective if applied properly. However, we believe that the optionality afforded to issuers, together with how banks set “price guidance” (discussed below) has the potential to result in predatory bidding strategies, which fuels inefficient price discovery.
Price guidance in an auction is a spread range that is provided by the arranging bank, assisting companies in issuing the debt and setting up the auction. This range uses many inputs, one of which is the previous comparative auction’s clearing spreads. The purpose of the price guidance is to provide some indication to prospective bidders of where issuers intend the clearing spread to be, as well as purporting to be a yardstick for a fair market value of the credit spread.
Given how bids are allocated in a Dutch auction, there is an incentive for investors to try to ‘undercut’ the market by submitting smaller volume, very aggressive bids (lower spread relative to price guidance) with the hope that the auction eventually clears at a higher spread, and they will enjoy that same spread and a high probability of being issued some of the bonds. In general, this practice can be dangerous in Dutch auctions, as it could result in the so-called “winner’s curse”, where it is possible that you win the auction at that aggressive price thereby “overpaying”. However, for the reasons we outline in this document, there are further dynamics at play which limit the probability of the winner’s curse occurring, making undercutting an attractive option for bidders and an exploitable feature of the auction process.
An example of undercutting is shown below with the BNP Paribas auction that took place in November 2024. The issuer was looking to raise R500m across 3-year (RCS01) and 5-year bonds (RCS02). The distribution of bid value to spread is shown below.
We can see that although price guidance for this bond was 135 basis points (bps) - 145bps, some investors bid as low as 90bps (in small volume) with the hope that it would clear higher, and it did.
The same can be seen in the 5-year bond where some investors bid small volumes as low as 120bps, despite price guidance being 185bps - 200bps.
Given the lack of transparency in terms of pricing and trading activity, combined with the relatively poor liquidity in the credit markets, most investors tend to be anchored to this guidance when bidding in auctions. Price guidance can be useful in providing insight to investors but over-reliance on this guidance, as an indicator of fair value, can lead to inefficient outcomes if it is not representative of fair value. Furthermore, issuers can control the clearing spread to some extent by using certain optionality afforded to them. The Equites Property Fund auction that took place in November 2024 clearly illustrates this. Equites came to market and gave themselves the option to raise between R500m and R750m. In other words, there was no fixed pre-determined amount. Furthermore, issuers are not obliged to issue anything if they choose not to. In this example, Equites were raising funding across a 3-year and a 5-year bond and did not commit to issuing a fixed amount in each bond. The outcome and bid distributions are shown below.
It is no coincidence that both these bonds cleared at the bottom end of price guidance. Equites chose to issue R558m bonds and despite receiving R825m in bids for the 3-year bond, only allocated R188m here and the remaining R370m in the 5-year bond. Equites was looking for up to R750m and received total bids of R1.7bn. Had they issued R750m equally across these bonds, the 3-year bond would have cleared at 120bps (upper end of guidance). Furthermore, there was sufficient appetite for the entire R750m to be issued in one bond. Had that happened, the 3-year would have cleared at 136bps (26bps higher than issued spread) and the 5-year at 138bps (13bps higher than issued spread). Therefore, there is a very wide range of possible clearing spread outcomes, all of which are determined by the volume that an issuer ultimately elects to issue after the auction bidding has been completed.
Generally, the market will anchor itself around the issued spread on the assumption it is fair value. In our view, the clearing spread is not necessarily a true reflection of fair value. By affording issuers with so much optionality, issuers could theoretically exploit the auction methodology to place downward pressure on clearing spreads. Furthermore, this clearing spread will be used to determine the next auction’s price guidance, which then exacerbates this problem in future.
We believe to get to the intended benefits of a Dutch auction, and to ensure equitable treatment of investors and issuers, issuers should commit to issuing a fixed amount per bond on auction and be compelled to do so if there are sufficient bids. To protect issuers, there can be a maximum spread set above which the issuer is not obliged to accept bids.
An alternative type of auction which could achieve similar results is called a Competitive Auction. This type of auction is not used in South Africa but is used in other countries. Namibia issues government bonds using this type of auction, for example. Here, like a Dutch Auction, investors bid for their desired volume and spread in a sealed-bid process. The difference however is the issuer will then accept or reject each bid at that specific spread. Successful bids will then each be issued at that respective spread, and not all at the same clearing spread as is the case with a Dutch auction. This means that there will be multiple issue spreads for the same bond. This can become burdensome from an administration perspective but does force investors to bid where they see fair value, thereby eliminating undercutting.
We have continued to bid in auctions at spreads that reflect the risks of investment, and where possible, have moved toward private placements rather than public auctions to secure high-quality assets at risk-reflective spreads. Investors need to consider the undercutting techniques used by some market participants, the setting of price guidance, and the allocation practices of issuers when they analyse public auction outcomes as indicators of fair value. The use of the results as inputs into future bidding strategies may imply that a bond’s spread is not reflective of all the underlying risks.
While local credit spreads have been tightening, this trend is not solely a reflection of risk reduction but is also influenced by supply constraints, growing demand, and auction mechanics. The Dutch auction process, coupled with investor bidding strategies and reliance on price guidance, can result in credit spreads deviating from fair value. Investors should be mindful of these dynamics to ensure informed decision-making in the evolving local debt capital market.
Source: Auction Outcome RCS BNP Paribas Personal Finance South Africa Limited, 28 November 2024; Auction Outcome Equites Property Fund Limited Senior Unsecured Notes, 11 November 2024.