Skip to content

On lending to the SA Government

Traditionally, government bonds are referred to as ‘risk-free’ assets. Much of modern financial theory, and indeed, the practical day-to-day workings of financial markets, depend on such a concept.

Peter van der Ross

Peter van der Ross

Portfolio Manager, STANLIB Absolute Returns

We price assets based on the additional (over and above risk-free) risk premia they expose us to so that over time, in order to entice investors, assets need to pay returns commensurate with their inherent risks. That’s the theory.
In reality, at STANLIB Absolute Returns, we do not regard anything as being risk-free, not even cash. Every asset brings with it a set of risks.


The reason government bonds are theoretically considered risk-free is that the government can simply print the currency required to settle its obligations. If the pace of money printing broadly matches the pace of income (GDP) growth, this is not a harmful practice, but it’s easy to see how any regime can be tempted to abuse the privilege of owning the printing press! Most governments give in to this temptation and as a result, global sovereign debt levels are at all-time highs, as even cash in circulation is an obligation, or debt, of the issuing central bank.


With the yield on many developed market government bonds being negative, professional investors have recognised the apparent shift from these markets offering risk-free returns to now offering return-free risks. The South African government bond (SAGB) market is something of an outlier, with prospective returns seemingly well in excess of inflation. However, SAGB’s bring with them significant and growing risks.


Many institutional money managers in South Africa have a predominantly long-term, valuation-based perspective. However, there is clear and present danger looming in the SAGB market. Short-term risks need to be navigated if we are to safeguard the long-term purchasing power of our clients’ savings. These short-term risks are concentrated around the scale and pace of our beloved country’s fiscal decay – SA continues to spend at an alarming rate, continually assuming there’s more where that came from. There is more – markets will almost always provide funding – but at what price?


At STANLIB Absolute Returns we assess the attractiveness of assets over a tactical (up to 12 months) horizon through six lenses: valuation, economics, liquidity, momentum, volatility and sentiment. In this paper we will evaluate the SAGB market through these lenses.

Valuation: in the eye of the beholder

A good long-term rule of thumb is that long-bond (10-year) yields tend to oscillate around GDP growth rates. However, this is a very long-term guide and many bond markets deviate from this for years at a time, with such deviations reflecting some or other distortion in the system. In SA, GDP growth is currently around 5% nominal, whereas our 10-year bond yields around 9%. On this basis, SAGBs are c.4 percentage points too high. Although this is a fairly crude measure, it gives us a useful starting point for assessing value. A more accurate valuation assessment requires us to build up the major components of the bond yield.



Any entity that is indebted has three options to repair its balance sheet:

  1. Repay the debt. This requires a growing level of income (tax receipts in the case of sovereign borrowers) to repay the capital and interest outstanding. If income growth is too slow, or non-existent, then…
  2. Default on the debt – this can be outright, or through a ‘re-negotiation’ of terms. This is a key risk for bondholders, so we use credit default swap (CDS) pricing to infer the price of insuring against default.
  3. Default by stealth, i.e. inflate the debt away. This works by borrowing a rand today and engineering inflation so as to ultimately repay that debt with a less-valuable rand. In this sense, inflation is the friend of politicians who struggle to take difficult or unpopular decisions, but it is the enemy of the investor in nominal bonds. (As an aside, governments that issue both nominal and inflation-linked bonds (ILBs) face a dilemma of their own making as inflation-linked debt cannot be inflated away!) Inflation is the path of least resistance for policymakers, especially in low-growth economies.

Given that context, we build up the components of the bond yield with a view to assessing whether, as bond investors, we are being adequately compensated for default and inflation risks.


We start with the yield on 10-year US Treasury (UST) bonds as, in a world where the US dollar is the de facto global currency, 10-year USTs are as close as we come to risk-free financial assets. On an objective build-up, USTs look expensive (an essentially flat yield curve that is roughly in line with US CPI and well below the US GDP growth rate), but for the sake of this analysis we take the spot rate as given:


Value is in the eye of the beholder because assumptions need to be made. Of the numbers above, the most difficult one to assess is the residual (additional) 3.4% premium that the SA Government, off spot inputs, is having to pay 10-year SAGB investors. The remaining elements can be sense-checked more easily.

Although the median of SA CPI since the introduction of inflation targeting in February 2000 is 5.3%, today’s Reserve Bank (SARB) reaction function clearly shows intent to bring inflation closer to 4.5%. Since a large portion of the CPI basket comprises administered prices, it’s going to be hard to sustain an inflation rate of 4.5%, unless the SARB crushes the demand side of the economy through keeping policy rates too tight. In our bond yield build-up, therefore, we could assume SA CPI of closer to 5%, signalling that forward-looking inflation is likely to be lower than in the past, but perhaps not as low as the SARB would like.

Conversely, the US Federal Reserve (Fed) has consistently undershot its 2% inflation target for the last decade, so we may be overestimating that rate. In other words, the inflation differential could be closer to 3% than the 2.3% shown above, implying that the market is not demanding quite as much as 3.4% per annum as an additional risk premium. (Note, there are technicalities around which measure of US CPI the Fed targets, but it remains accurate to say that for the last decade or so, the Fed has been working hard to engineer higher, not lower, actual and expected inflation).
Although CDS pricing is observable, there are some wrinkles with including it in the manner in which we have. Firstly, 5yr CDS are much more liquid than 10yr CDS, so better 5yr price discovery gives us a more reliable input, even though we’re trying to assess a 10-year bond. Second, credit default swaps represent the cost to a global investor of insuring against an issuer defaulting on its US dollar-denominated debt. When the chips are really down and some sort of explicit default is inevitable, the probabilities of default on rand- versus dollar-denominated debt will not be the same. As a real-time approximation of default risk, though, it’s the best we have.

The following chart shows the movement of the SAGB component risk premia over time:

Source: Bloomberg, STANLIB Absolute Returns calculations

Before 2013, the residual item (or the additional risk premium that the market demands) moved around mostly because our inflation volatility – in green – used to be higher. At times, the market looked beyond cyclically high or low inflation, however, our contention is that this time SA inflation is low due to secular domestic stagnation and relatively tight monetary policy. The market doesn’t seem to believe that inflation will remain around the 4-4.5% level.
Given that SAGBs are currently offering c.3.4% more yield than they theoretically should, how do we make sense of this market pricing? Implied in market pricing is either that:


  1. All else being equal, inflation over the next 10 years could average 7.5-8% (being spot inflation plus 3.4%). This would be a shock to markets but is consistent with the argument of the political path of least resistance.
  2. SA’s probability of default is more in line with somewhere like Turkey, whose 5yr CDS trades at 270bps versus SA at 170bps, and SA inflation will be in the region of 6.5-7%. (In other words, of the 3.4% residual item, 1% could be for heightened default risk – SA CDS is mispriced – and the remaining 2.4% is compensation for heightened long-term inflation risk). Bear in mind that Turkey is rated three notches lower than SA by ratings agency Moody’s and two notches lower by Fitch and S&P.

Although these default and inflation scenarios are certainly not beyond the realm of possibility, they help us quantify the magnitude of the bad news that is priced in. On balance, we conclude that the buffer is sufficient to make the asset look cheap in absolute terms. SAGB’s are already fairly priced for those scenarios. We would have to believe that our future looks worse than those scenarios in order for the asset to look expensive.

Of course, assets don’t trade in a vacuum and with capital being globally mobile, we think about both the absolute and relative value they offer. For context, we plot the level of real bond yields (nominal yield less an average of inflation in the last two years inflation to avoid one-year base effects on CPI in some instances) across a collection of global developed (DM) and emerging markets (EM):

Source: Bloomberg, STANLIB Absolute Returns calculations


Generally, real yields are higher in EM than in DM countries. The z-scores – green bars in the chart above – show that many DM markets are one to two standard deviations more expensive than they’ve been since 2012. This creates a powerful push factor for capital to leave those domestic markets and travel the globe, searching for higher yields and, in the process, often driving those yields lower. Despite this push factor, foreigners own a smaller share of the SA bond market than they did two years ago. We look cheap, but are being shunned by global investors.



SA’s super-long 20-30-year bonds have sold off even more than the benchmark 10-year point, with the spread between SAGBs due in 2044 (R2044) and 2026 (R186) at all-time wides, as can be seen in the following chart:

Source: Bloomberg


Comparing a similar group of global bond markets at the 20-year point on their yield curves shows a similar picture, with a more extreme result:

Source: Bloomberg, STANLIB Absolute Returns calculations


At 5.5%, the real yield on 20-year SAGBs is approaching the long-term real return achieved on SA equities at c.7%. Using spot CPI, the real yield on 20-year SAGBs is closer to 6.5%! There is clearly a lot of bad news priced into the SA bond market as it has already discounted some sort of unpleasant economic scenario to come.


Valuation conclusion: Attractive. We favour longer dates with equity-like real return prospects.


Economics: buy now, pay later

The bad news being priced in is that SA’s national debt is simply unsustainable and so far, indications are that there is insufficient willingness or ability to arrest the decline. The profile of SA’s government debt to GDP since the 1970s is as follows:

Source: Bloomberg

The national debt burden is rising at an accelerating pace because of:

  • Persistently high government expenditure, despite talking tough around reining it in
  • Where there has been expenditure, it has not been directed towards growing or even adequately maintaining the nation’s capital stock, so there is not much prospect of a return on investment
  • A stagnant economy where the annual budget deficit including debt service costs exceeds the nominal rate of GDP growth (more debt, divided by the same level of income out of which to service it).
  • The steadily rising cost of servicing that debt, as both the stock of debt and the level of long-dated interest rates (bond yields) have risen.
  • Importantly, the chart excludes state-owned entity (SOE) debt, contingent liabilities and hidden liabilities (maintenance of roads, water infrastructure, etc).

National Treasury holds weekly auctions for both nominal SAGBs and ILBs. The profile of the value of the rolling three-month issuance of nominal bonds is as follows:

Source: ABSA Research


The Medium-Term Budget Policy Statement in October 2018 seems to have been the watershed moment when National Treasury acknowledged that the borrowing requirement was ramping up dramatically. Since then, the increased supply of SAGBs has caused a marked steepening in the yield curve, as market participants have re-priced their lending terms upwards. As we said up-front: the market will almost always provide funding, but at what price?
Interestingly, in 2012 ILB issuance was as high as 25% of total weekly issuance, whereas latterly that figure has dropped to around 15%. Again, this gently nudges the door open to a bit more inflation – all else being equal – in the long term.

The last national budget speech delivered by Minister Mboweni was brutally honest in its portrayal of the fiscal trajectory. Since then, Eskom has required greater-than-budgeted cash injections, tax receipts have disappointed materially, irregular expenditures have ballooned, and job losses have accelerated. As bad as the minister told us it would be last February, later this month he will have to admit that the country has further underperformed.

The rising debt/GDP trend is well in force and as a nation SA will battle to stop it breaching 90% in the next few years. The reason seems to be that the political landscape is too complex to allow for difficult decisions to be taken with the requisite level of urgency. Many people will differ with this point, but the reality is that financial market participants are likely to lose patience before the voting public does and the level of yields and steepness of the SA yield curve suggest that investors are already voting with their feet. Even with an SOE as relatively (compared to Eskom) inconsequential as SAA, there has been much tough talk, but very little follow through. This newsflash hit our screens recently:

We attach much more weight to actions than to speeches. While we fully appreciate the delicate nature of the balance of power, the negative inertia that set in during the Zuma years requires drastic, decisive action to reverse. Unfortunately, a precarious balance of power does not support decisiveness. The market is probably correct to discount an unpleasant economic scenario, and if the current trend is not arrested, SA is likely to be downgraded by more than one notch into junk territory.

Economics conclusion: Neutral. There are competing forces of increased issuance and very low tactical inflationary pressures at play.

Liquidity: A rising tide lifts all boats
The key consideration here is global liquidity, otherwise referred to as ‘global financial conditions’. Generally, risk assets do well when liquidity is plentiful and vice versa. Because EM sovereign bonds offer higher yields than their DM counterparts, their carry (or additional yield pickup) makes them typical beneficiaries of portfolio inflows when global financial conditions are easy. In other words, when the opportunity cost of domestic DM capital is low, as it is now, DM investors are forced to seek returns elsewhere. In the global context, therefore, we regard EM and SA bonds as risk assets, as they attract inflows when global investors are in risk-seeking mode and are sold when global risk aversion spikes.

The following chart demonstrates that when global financial conditions improve, EM bond spreads tend to contract. In other words, EM bonds do well when global financial conditions improve.

Source: Bloomberg


Global liquidity conditions are currently very supportive of high carry instruments such as SA nominal bonds. Courtesy of the Fed tightening monetary policy in 2017 and 2018, and President Trump’s trade wars, 2019 saw a slide in global trade volumes, business confidence levels and manufacturing PMI’s that, during the second half of the year, sparked the largest concerted monetary policy easing across global central banks since the 2008 Global Financial Crisis.


Both high-yield and investment-grade corporate bonds have attracted massive inflows as thin-to-negative DM sovereign yields have continued to force investors down the risk curve. EM sovereign bond yields typically trade in line with DM high-yield corporate bonds, so as long as the currency risk can be managed (or accepted), they are a natural potential beneficiary of flows. Again, if we plot global sovereign bond yields against where they were a year ago (in the following chart), most yields have compressed, except for SA. Note that we’re not arguing for SA yields to catch up, but the global search for yield will support SA yields from selling off much further.

Liquidity conclusion: Attractive. Global dynamics favour the carry-attractiveness of SAGB’s.


Momentum: The trend is your friend
“As long as the music is playing, you’ve got to get up and dance.” – Charles Prince, former Chairman and CEO of Citigroup

We deliberately apply a momentum lens to our tactical thinking in order to balance, in particular, the more qualitative valuation-based considerations with more dispassionate, quantitative inputs. This element of the process is designed to keep us in markets that are doing well and help us avoid markets that are doing poorly, despite what we might think about those markets.

We observe the overall returns to bonds versus cash. Recall that the total returns to bonds are the sum of capital and income returns. The (currently) much higher yields on bonds mean that there is some buffer for small capital losses on bonds before they return less than a fair, or average, term premium in excess of cash. The following chart shows rolling 52-week excess returns on SAGBs versus cash. Bonds have beaten cash by, on average, 2.3% p.a. since 2011, but that excess has recently fallen to 1.5% – a poor momentum signal, but only just.

Source: Bloomberg, STANLIB Absolute Returns calculations


Absolute SAGB yield trends: flat at the 10-year point, but negative momentum in longer-dated yields (next chart):

Source: Bloomberg, STANLIB Absolute Returns calculations

Momentum conclusion: Neutral at the 10-year point; Unattractive for longer-dates.

Volatility: Here today, gone tomorrow
“Stability begets instability.” – Hyman Minsky

Whereas we want to position in line with market momentum, we use volatility as a countercyclical indicator. A volatile market is likely to be selling off, prompting a buy signal. A quiet market is likely to be complacent and at risk of a selloff.

In the following chart we plot rolling three-month realized volatility in the ALBI and USDZAR, overlaying with the 5yr CDS to show spikes in risk aversion. We clearly see heightened risk aversion around the 2008 crisis, as well as in 2016 when global equity earnings growth was negative year-on-year, coinciding with (but unrelated to) the unexpected firing of Finance Minister Nene. We find that SAGB volatility is currently at around the lowest level it has been since 2008!

Source: Bloomberg, STANLIB Absolute Returns calculations


A complacent market is a dangerous market, so we use our volatility lens to nudge us towards selling such markets.
It is important to note that markets do gradually change over time. Since the turn of this century, the duration or interest-rate sensitivity of the ALBI is up by 55%, from 4.5years to 7 years. This has been a function of National Treasury issuing longer-dated stock. (The Treasury does seem to be somewhat alive to the greater interest cost it is incurring, and over the last year or so has issued stock more in the belly of the curve, reducing ALBI duration from its peak of 7.5 years). ALBI duration over time:

Source: JSE

Although ALBI duration is up by 55%, inflation and hence interest rates are much less volatile. The two effects have largely netted off, leaving overall ALBI unchanged over the period. However, given the heightened risk of higher inflation going forward, we must be aware of the fact that ALBI volatility now has the potential to rise much higher than in the past.

Volatility conclusion: Unattractive. Price action suggests a complacent market.


Sentiment: Buy on the cannons, sell on the trumpets
We also use sentiment as a contrary indicator in our framework. The logic is that where expectations around an asset or asset class are high, that market becomes more likely to fall short, disappoint, and lose the investor money, and vice versa. This is not a given, of course, but when investing we are always dealing with probabilities with the aim of constructing portfolios that maximise our probability of success.

We have liked SAGBs since 2016, for a number of reasons:

  • Our then strong rand view and bearish view on the SA economy meant that inflation was likely to stay low (note that we no longer hold a strong rand view).
  • The yield curve had begun steepening, so real yields on longer dates were very attractive.
  • We were bearish on SA equity, so SAGB’s could form a relatively larger part of the portfolio.
  • The global backdrop favoured carry trades.

Our favourable outlook on bonds has worked for our clients but as we stand today, we observe that that has become the consensus view among SA fund managers.

Bank of America Merrill Lynch (BofA) surveys asset managers every month, and its results are amongst the multiple inputs we categorize under sentiment. Some results from recent surveys include:

  • The percentage of SA fund managers who thought bonds were overvalued was roughly 25% in November, 15% in December, and zero in January.
  • In January a record number (67%) of managers thought bonds were cheap
  • For the first time since 2000 bonds are the favoured domestic asset class.
  • Despite this, only 20% see a financial solution to Eskom and only 27% see an operational solution.
  • Regarding the upcoming budget speech, a net 33% expect the budget to be market negative, 60% say it will be either outright positive or that the bad news is all in the price, so the event won’t be market negative.

Reading those positioning and outlook summaries together with very low volatility at present, we would conclude that sentiment regarding SAGB’s is complacent, so we should proceed with caution.

Sentiment conclusion: Unattractive; becoming a relatively crowded trade domestically.

So, what to do?
This framework helps us think clearly around how, and at which levels, to position in assets. We do not necessarily equal-weight the above signals and need to be mindful that:

  1. Overall portfolio positioning is also a function of our views, based on similar analysis, of competing asset classes.
  2. Different mandate requirements amongst the various portfolios under our stewardship will see us expressing the same asset class views and concerns to a greater or lesser extent.
  3. The signal from some factors like volatility and sentiment can swing much faster than others, like economics, liquidity and valuation.

Our conclusions per lens are:

Valuation: Attractive. We favour longer dates with equity-like real return prospects.

Economics: Neutral. There are competing forces of increased issuance and very low tactical inflationary pressures at play.

Liquidity: Attractive. Global dynamics favour the carry-attractiveness of SAGB’s.

Momentum: Neutral at the 10-year point; Unattractive for longer-dates.

Volatility: Unattractive. Price action suggests a complacent market.

Sentiment: Unattractive; becoming a relatively crowded trade domestically.

Clearly, although SAGBs look generationally cheap, the risks within the asset class are building, with a couple of our indicators having turned negative very recently. With mounting event risks around the budget speech, the looming Moody’s event (are we there yet?) and SOE issues being front and centre, a more cautious approach to SAGB’s seems warranted.

On balance, over our 12-month tactical horizon, we remain positively predisposed to SAGBs. However, given the mounting event risks we would look to risk-manage these exposures through – depending on market pricing – some combination of shortening duration, implementing outright hedges on SAGB’s and remaining relatively fully exposed to offshore assets.

More insights