Are SA bonds still attractive?
Our SA bond-positive macro narrative since August 2023 has been that, although the country faces structural risks that are likely to lead to a downgrade to single B over the long run, the global monetary policy cycle is bond positive. This narrative was supported by the US Federal Reserve’s (Fed’s) halting of interest rate increases in November 2023 and the sustained inversion of global developed market yield curves.
However, as is often the case when economies reach a turning point, recent data releases provide a far more complex picture. While inflation is slowing, it remains above target and growth appears to be relatively resilient in the face of restrictive financial conditions; US unemployment remains below 4% and, although it has risen from 3.3% to 3.7% it is nowhere near recessionary. Supply-side inflation risks are also rising due to elevated global shipping costs and geopolitics is being cited as one of the risks identified for 2024.
US real bond curve is almost flat, pricing out a soft landing.
The nominal bond curve (2-yr versus 10-yr) is now only marginally inverted (-36 basis points (bp)), and the real curve is almost flat. However, it remains unclear what will continue to drive dis-inversion. Will the US enter a recession (hard landing) resulting in the curve bull flattening? Will inflation slow to 2% allowing for 150bp-200bp of Fed rate cuts (soft landing)? Or will inflation remain above target at around 3.0% resulting in either fewer cuts, i.e. 50bp, or no cuts?
Over the past few weeks, and specifically after the US January CPI data release, consensus has started to move towards the latter narrative. The back end of the yield curve has risen on higher inflation expectations and the front end is pricing a shallower cutting cycle with interest rate derivatives pricing for 100bp of cuts by the end of the year, versus 150bp a month ago.
Compared with January, the US Treasury (UST) 2-yr yield rose 44bp, to 4.64% and the UST 10-yr rose 37bp, to 4.28%. The US 10-yr TIP (US real rate) rose 28bp, to 1.9%, while the 2-yr TIP rose 11bp to 2.06%. The real curve in the US is now almost flat, implying that markets expect real rates to remain where they are.
A no landing narrative could see 10-yr UST at 5.8%
The no landing narrative argues that US inflation and rates have normalised to pre-GFC (Great Financial Crisis) levels, which is true, and that the aberration that was negative real rates between 2008 and 2022 is now over, which may be true. Prior to the GFC the average spread between the 2-yr and 10-yr Treasury was 83bp. If this narrative plays out, and the 2-yr UST moves above 5.0%, on inflation fluctuating around 3.0%, then the 10-yr yield could rise to 5.8%.
SA rates will be determined by US CPI
US CPI will determine the outlook for SA rates, which requires that we unpack US CPI in more detail.
In January, US CPI fell to 3.1% year-on-year (y/y) from 3.4% in December but was higher than the expected 2.9%. The upside surprise was driven by services prices, as core goods fell for the eighth consecutive month. Services inflation (ex-energy) rose 0.7% month-on-month (m/m), the most in 18 months, keeping annual services inflation elevated at 5.4% y/y. The culprit within services is shelter prices (rental inflation), which rose 6.0% y/y and has a significant weight of 36.2% in the Index. Excluding shelter, services CPI was a more acceptable 3.6% y/y in January.
What is keeping rentals elevated and will it remain elevated? The data so far suggests that rental inflation will continue to moderate. Firstly, rental inflation usually peaks three quarters after the peak in headline inflation and this cycle is no different (headline CPI peaked in mid-2022 and rental inflation in 2Q2023.) So, the current lagged response in rental inflation is usual. In addition, the CPI rental inflation index measures the change in average rentals paid, which includes historical rental agreements entered when inflation was higher.
The US Bureau of Labor Statistics (BLS) issues a quarterly New Tenant Rent Index which captures only new rentals. As evident in Graph 1, new rental inflation has fallen dramatically, from 2.6% y/y in 3Q23 to -4.6% y/y in 4Q23. The relationship between new rentals and average rentals is intuitive and implies that shelter inflation should moderate over the coming months.
Graph 1: New Rentals index leads Rental CPI Index
Graph 2: US CPI Shelter
Interestingly, looking at the SA vs US forward rate agreement (FRA) curves since the Fed pivot in November 2023 implies that the extent to which markets expect the SA Reserve Bank (SARB) to cut is driven more by SA’s inflation forecast than by expectations of how much the Fed will cut; however, timing, i.e. when the SARB is expected to start cutting, is driven more by the Fed’s timing.
The transmission mechanism from rate hikes to weaker demand
Economists are pondering a broader question about the effectiveness of interest rate hikes in the US. There is a compelling argument to be made that this cycle is different because the Fed has less control over inflation than in the past. Firstly, an estimated 90% of mortgage debt is fixed and corporate debt is long dated. In addition, the strength of house prices and the equity market has resulted in the wealth effect keeping demand strong in the US.
SA living budget-to-budget, election-to-election
South Africa’s revenue collection is currently in line with the MTBPS (Medium Term Budget Policy Statement). This isn’t as positive as it sounds, considering that revenue growth was budgeted to grow by only 1.02% in nominal terms and contract 4% in inflation-adjusted terms.
Expenditure, on the other hand, was significantly underestimated by Treasury. The MTBPS budgeted for nominal expenditure to grow by only 1.8% y/y. Adjusted for inflation, this would require that government reduce spending by 3.2%. This is unrealistic in a normal year and unattainable leading up to an election. The budgeted growth in spending was also naïve given that the wage agreement allows for a 5.1% annual increase, and debt service costs, which now account for 14.5% of spending, are growing at a whopping 17% y/y.
Year-to-date, total main budget spending is growing at 7.7% in nominal terms. If spending were to continue at this rate, overspending would total R119bn for the fiscal year! Overspending in 2023/24 is likely to be financed by drawing down on cash reserves by more than budgeted, meaning Treasury starts the 2024/25 year with less cash in the bank.
Given the relatively significant amount of time that the SARB has spent flagging drawing down on the gold and foreign exchange contingency reserve account (GFECRA), we anticipate that it will go ahead with this, which will be positive from a cost of funding perspective. Our understanding is that National Treasury will be able to borrow the unrealised gains from their gold and foreign exchange reserves at the repo rate. This is well below its current average borrowing costs, which are closer to 11%, and we expect that these funds will be utilised in tranches over the next few years to reduce reliance on government bond issuance.
Repo rate cuts to start later in 2024.
Consensus expectations that rate cutting in the US and SA will only start in mid-2024 means money market rates remain attractive (9% on 1-year NCDs) especially on a volatility adjusted basis. Implied real yields offered by nominal government bonds of 6% appear to be pricing in meaningful sovereign risk. Inflation-linked bonds have weakened to real yields of 5.0% which is also attractive, albeit resulting in inflation break-even of 6%-7%.