Liability driven investing in South Africa

04 August 2014

Since the issue of the first inflation-linked South African government bond in March 2000, the universe of inflation-linked bonds available in South Africa has expanded beyond just government bonds. As at the end of May 2014, there were nine issues of government inflation-linked bonds listed on the Johannesburg Stock Exchange (JSE) with a market value of R369.3bn.

This compares to ten non-government issuers (excluding special purpose vehicles and notes) of listed inflation-linked bonds with an outstanding market value of R72.1bn across 38 issues. Within the non-government issuers of inflation-linked bonds, the local banks are the largest issuers. Their total issuance of R38.3bn across 65 issues amounts to 53% of non-government issuance and exceeds that of even the State Owned Companies (SOCs). By comparison, the SOCs have issued inflation-linked bonds with a market value of R33.4bn across nine issues.

Liability driven investments

LDI portfolio managers have been buying bank-issued inflation-linked bonds introducing the additional element of credit risk to their de-risked portfolios. The credit skill and experience of the LDI manager therefore becomes an important criteria in the mandate granted to LDI portfolio managers.

Inflation-linked bonds (or linkers) have significantly different cashflows to their fixed rate counterparts.

Bond investors are, in effect, lending money to bond issuers. Investors expect this money to be repaid when the bond repays or redeems.

• At redemption the buyer of an inflation-linked bond is repaid their invested amount plus growth in line with inflation
• At redemption the buyer of a fixed rate bond is repaid their invested amount only.

If this were the full story then this would be an unfair deal for the buyer of a fixed rate bond. However, the fixed rate bond buyer typically receives a higher interest payment through the lifetime of the bond. This observation is important for a lender, especially if the investor has concerns about the creditworthiness of the borrower. Inflation-linked bondholders expect their invested capital to be returned also later than if they had invested in a comparable fixed-rate bond. Bond issuers understand this feature of inflation-linked bonds. It is a feature that implying that they effectively lend the bond issuer money for longer.

In the case of utilities or infrastructure providers, such as Eskom and The South African National Roads Agency Limited, generally known as SANRAL, issuing inflation-linked bonds to fund their borrowing requirements makes sense since their revenues are in part linked to inflation.

The fundamental case for banks issuing inflation-linked debt is less clear although banks are probably best placed to understand the risks and rewards of doing so. With sophisticated treasury teams, banks are able to hedge any risks posed to them from the issuance of such debt whilst ensuring that the favourable economics on offer, such as being able to capture greater borrowing for longer, is retained.

The chart above shows the total size of all non-government issuers’ inflation-linked bond programmes across all maturities and issues. Apart from Eskom, SANRAL and the Trans Caledon Tunnel Authority (TCTA) the other main issuers of inflation-linked bonds are the banks. The aggregate issuance by the banks is greater than the aggregate issuance of Eskom, SANRAL and the TCTA.

Bank issued inflation-linked bonds present a very different credit risk profile to their government and government-guaranteed counterparts and needs to be evaluated differently before inclusion in an LDI portfolio.

Clearly a thorough credit analysis must be carried out prior to investing in bank-issued inflation-linked bonds. More than that, the long-term nature of the funding being provided by the debt holder should be understood and properly compensated for.

This is critical in an LDI portfolio which is designed to be a hedging portfolio intended to mitigate risk. This should extend to mitigating risk in stressed market conditions.

A further risk is liquidity risk. Since bank issued bonds tend to be of a smaller size, trading in the secondary market is much more limited and bid-offer spreads (the costs of buying and selling) are higher than government and governmentguaranteed inflation-linked bonds.

The total size of bank issues (per issuer) range between 27% – 46% of the total issue size of government-guaranteed Eskom bonds. This feature, combined with the government guarantee on Eskom bonds, makes the latter far more liquid than bank-issued bonds.

Moody’s, in a report on the South African banking system, released in May 2014, noted that the outlook for the local banking sector remains negative – mainly the result of subdued economic growth, which will continue to exert pressure on asset quality. We have also seen recent negative rating action on some South African banks, and rating outlooks, in most cases, remain negative. In our view, this places any decision to hold non-government guaranteed bonds in the LDI portfolio firmly in the spotlight.

The investment decision to increase the pension fund’s strategic exposure to inflation-linked bonds, could, in our view, be separated from the decision to take credit risk…

Credit risk, in turn, should be further decomposed to ensure a proper diversification of idiosyncratic credit risk as well as some consideration of the asymmetric nature of credit risk.

For funds that are comfortable awarding their LDI manager a degree of flexibility then the LDI manager could access a ‘wider toolkit’ to enhance the yield on relatively lowyielding government inflation-linked bonds.

This ‘wider toolkit’ would include allowing the manager flexibility to synthetically create the exposure to the inflation-linked bond and use the cash to invest in a diverse portfolio of credit-risky bonds.

When investing in credit one of the commonly held tenets is to diversify idiosyncratic risk by ensuring exposure to a wide spectrum of issuers. By synthetically creating the inflation-linked bond exposure, the manager is free to diversify the credit-risky bond portfolio across different issuers (irrespective of whether the bond itself was inflation-linked) and in so doing avoid concentration risk to a single issuer.

However, the South African listed corporate bond market is concentrated and so the benefits of this diversification are limited. As an example, there are only 52 different companies with listed bonds. Even if we extend our definition of credit, beyond listed bonds, to include loans to companies that have been syndicated by the banks making those loans, our estimate is that the average South African investor is only able to access no more than 100 different South African companies. This is in stark contrast to the world’s most liquid corporate bond market, the USA, which has more than 600 different issuers.

“The unlisted market could, in our view, assist South African investors in their diversification efforts by offering access to at least a further 400 different companies.”

With the nascent state of our corporate bond markets, achieving further diversification requires investors to consider the unlisted debt space as an avenue for obtaining credit risk exposure to a wider spectrum of South African companies. The unlisted market could, in our view, assist South African investors in their diversification efforts by offering access to at least a further 400 different companies.

The second commonly held tenet is that any consideration of credit risk should include a proper evaluation of the asymmetric risk-return profile of corporate bonds. By ‘asymmetric’ we mean that, unlike equities, corporate bonds have limited upside and unlimited downside. ‘Limited upside’ means that the maximum annual return is known at the outset and is defined by the gross redemption yield on the bond. This is in contrast to equities where the potential for capital gain is, in theory, unlimited. But both corporate bonds and equities have ‘unlimited downside’. That is, as with an equity investment, investors in corporate bonds can lose their entire investment.

To properly account for the asymmetric nature of their investment in credit, investors should expect their manager to also consider the risks associated with default and the subsequent likelihood of recovering if not all, then at least some, of their initial investment. This then leads to the concept of secured credit versus unsecured credit.

Listed corporate bonds are generally unsecured in nature. Diversifying into unlisted credit can allow investors to access secured credit assets which, in turn, can offer investors greater protection should the credit worthiness of the borrower deteriorate.

It is a fallacy to believe that because an LDI portfolio’s holding in a credit-risky bank issue is only a small part of the overall LDI portfolio, that diversification of idiosyncratic credit risk has taken place.

A bond-only LDI portfolio may be stocked with government and government-guaranteed bonds. Adding a relatively small holding in a single name bank bond such as our example above may be considered prudent since the holding is small relative to the total LDI portfolio. Is this really the case?

In our view, a more appropriate way of evaluating the holding would be as set out below:

• For the pension fund to consider what total exposure it would like to credit as an asset class (through a risk budgeting and strategic asset allocation exercise which allows for differences in the risk-return profile of different credit instruments e.g. listed versus unlisted; secured versus unsecured).
• To then control for idiosyncratic risk by setting appropriate parameters such as issuer, issue and sector concentration limits.

Not to follow such an approach raises the risk that the pension fund can end up overexposed to a specific issuer or sector at portfolio level. This could arise since the fund’s other asset managers (i.e. those not managing the LDI portfolio) could also be exposed to the same creditrisky issuer through their portfolios which hold their equity or indeed their bonds.

When drafting the investment guidelines for an LDI mandate, clients should set explicit parameters around the types of inflation-linked bonds that the manager is permitted to include. For managers with appropriately skilled credit teams, allowing the manager to buy credit-risky inflation linked bonds could enhance the yield on the LDI portfolio. However, even then clear limits should be agreed with the LDI manager and those limits should be set based on the pension fund’s risk budget.

Ashburton Investments’ Solutions team comprises internationally experienced LDI managers who have managed some of the largest LDI mandates in the world across the UK, USA and now South Africa. We also have a deeply experienced credit team.