Wednesday 19 May 2021 by Andrew Mayes Bond-sell-off-takes-a-pause-800 Company research

Bond sell-off takes a pause

Inflation expectations move into target range as participants debate whether inflation is more than just ‘base effects’ and other transitory factors.

Over the next year or so it is almost certain we will see episodes of rising prices as the weak economic data of 2020 is replaced by more ‘normalised’ data in annual measures of inflation (the ‘base effects’) (see Figure 1), despite the unambiguously soft inflation pressures reported in the most recent March quarter.


At the same time, certain pressure points stemming from a demand-supply mismatch are already emerging. As widely anticipated, an inflationary release of pent-up demand has emerged (aided by unprecedented fiscal and monetary support) as restrictions on movement ease, and supply chain disruptions have subsequently ensued--supplier lead times in some parts of the world have lengthened to the greatest extent on record. The prices of many commodities have surged to levels not seen for years, noting that temporary pandemic-related supply interruptions are as much to blame as strong forward demand.

Input prices into manufacturing can serve as a leading inflation indicator (among others); as producers are required to pay more to suppliers, the cost of the final product will invariably grow (see Figure 2). So-called cost-push inflation (rising prices from rising input costs) is normally temporary and generally looked through by central banks.

The question is whether price pressures are simply a reflection of catch-up growth or set to become more entrenched.


Attend a FIIG Event

For now, markets appear to be on a similar page to current central bank thinking. Breakeven inflation--a market indicator of the level of inflation over a period of time, equal to the difference between nominal and real yields--peaks over the next two years before drifting lower, but still comfortably within central bank targets (see Figure 3).

Given easing restrictions on movement and high levels of unemployed (relative to pre-pandemic levels), in time we would expect an increase in the availability of labour and other components in the supply-chain to relieve many of these constraints (and associated inflation)--but it’s unlikely to be a smooth transition, and some inflation may just turn out to be more than just transitory. With central banks targeting actual inflation--not merely expectations or forecasts--and willing to maintain a supportive policy until their objectives are met, the risk to the upside of higher inflation is likely to remain elevated for much of this year.

Moreover, households are still sitting on higher levels of savings relative to pre-pandemic levels, which are likely to be spent at some point in the future. However, with much of that build-up in savings a function of restrictions on movement or government stimulus (such as wage subsidies), the subsequent growth outcomes feel more like a supplemented catch-up rather than a step change reflective of a stronger consumer.

Headwinds to a structural shift in inflation are likely to still persist

Key to meeting central bank objectives of full employment and inflation will be to close out the output gap (the difference between actual output and potential output) and reduce underemployment--a key impediment to stronger (wage) inflation over the last decade in Australia.

To do so would be to overcome the moderation of inflation over past few decades, influenced by an increase in the supply of capital (a function of a number of variables, including demographic trends and the fungibility of said capital) and reduced demand for capital. Even with record low interest rates, non-mining investment remains uncomfortably low as a percentage of total output. There is little to suggest current monetary policy settings will overcome this.

As the unemployment rate declines, the availability of suitable labour also declines (a higher number in Figure 4). As the demand for labour remains unchanged (or indeed increases) and supply decreases, the number of underemployed individuals’ (those who want to work more hours) should also decrease.


Australia’s underemployment rate has been strongly correlated with wage growth over the last couple of decades (see Figure 5). The central bank has also recently suggested the theoretical level of unemployment below which inflation would be expected to rise (and with it, stronger wage growth) may be less than 4% (from 4.5% pre-pandemic), or around a similar level to when we last witnessed decent wage growth (see Figure 5).

It is conceivable the closure of Australia’s international borders--in-effect reducing the supply of capital--will support the RBA’s ability to reach full employment and generate stronger wage inflation which has evaded it for the past decade or so. More realistically, and while we believe a return to pre-pandemic status quo is unlikely, we feel that as pressure on wages start to emerge, the political response will revert to the mean and an increase in the supply of capital (labour) will ensue--particularly if improving vaccination rates pave the way for greater freedom of movement.

While we would be more than happy to be wrong, we stick with our view that monetary policy alone is unlikely to drive down unemployment (and more pointedly, underemployment) to levels sufficient to generate wage growth above 3%. The question then is whether the central bank will continue to maintain its current policy settings out to 2024.

Bond sell-off has likely accounted for the recent developments, but we see room for higher yields as the economic recovery progresses.

Given the level of fiscal support globally, particularly in the US, we see room for yields to trend higher this year as the economic recovery progresses--although we expect the increase to be slower with the snap-back to pre-pandemic levels and sell-off on concerns over premature tapering largely behind us. While the recent move up in longer-dated yields may have been a bit too much, too soon (see Figure 6), with the RBA anchoring short-term rates, we feel the overall direction in longer-dated yields is likely to remain higher.


After all, even at 2%, long-term bond yields are not particularly high in any real sense (see Figure 7).


Longer-term yields are mainly a function of economic growth and inflation expectations, as well as the supply and demand of longer-maturity government bonds. Even without rising inflation expectations (which appear to have settled for now within central bank targets), yields can continue to rise. With real yields--a proxy for growth expectations--still in negative territory (see Figure 8), there is room for them to revert back into positive territory longer term if the economy’s prospects continue to strengthen.

One approach (again, among others) when assessing the outlook for growth is to look at the ratio of copper to gold. Because copper is a key industrial metal that is used globally in a wide range of industrial applications, it performs strongly when the global economy is similarly performing strongly (gold is the opposite). As such, it can provide a useful indication of global growth and inflation expectations, and has done in recent history (see Figure 9).

As noted earlier, the price of many commodities has surged to levels not seen for years. The question is whether the recent decoupling between the ratio of copper/gold and longer-dated yields is a reflection of the forementioned transitory factors, or something more structural (the use of copper as an input into green energy generation, without a corresponding increase in supply, may also be contributing to a higher base going forward).


For now, we think the demand-supply mismatch explains the divergence. At the same time, as individuals emerge further from lockdown and a greater sense of normality resumes, stronger demand for services will emerge, which should detract from some of the growth in consumption of goods (but should otherwise preserve the total level of aggregate demand). We therefore think convergence between the two will take place somewhere in the middle (i.e., higher nominal yields, but less than what the chart may otherwise indicate).

There is also the added complexity of central bank intervention and the risk of an unanticipated tapering of bond purchases leading to further volatility (see Figure 6) (recall longer-term yields are also a function of the supply and demand of longer-maturity government bonds). Central banks are likely to be less concerned about the quantum but more about reason yields are rising, and what effect that is having. In Australia, this particularly concerns financial conditions (shorter-end) and the currency (longer-end). With financial conditions very accommodative, the RBA continues to purchase longer-dated government bonds (QE) (which are in-effect about 30 basis points lower by their estimates) in an effort to sustain a lower exchange rate than otherwise (see Figure 10 and 11).


The old adage ‘don’t fight the Fed’ seems fitting at this point. While we expect yields to move higher, as long as the central bank remains an active purchaser of government bonds, which appears likely once the current program of bond purchases is concluded in September, the rise is likely to be tempered.

While much of the focus has centered on the US, as they are likely to be the main driver of global yields, some countries have already taken a step back, including Canada--which was widely anticipated and had little bearing on yields. Whether the eventual taper triggers a volatile sell-off comparable to that witnessed earlier in the year or a gentle glide higher depends on the signalling from central banks, including the RBA.

Fixed Income Implications

The exit path from unprecedented support is likely to be complicated and volatile. With central banks willing to tolerate higher inflation, any sense that stronger inflation carries more than a transitory effect is likely to see markets respond sharply (wider yields). Clearly, the reverse is also true. As we indicated above, on balance we believe yields are more likely to rise than fall this year, but we are unlikely to witness a similar magnitude to that experienced earlier in the year.

For bondholders, there is no shortage of variables to consider. We would offer three overarching considerations given the current market dynamics:

  • Don’t be too hasty in shying away from duration. Longer-dated government bond yields have been declining for more than 30 years, only briefly interrupted by periods of selling (see Figure 7). Central banks are likely to remain active until substantial progress toward their objectives are met, which should supress yields (support prices). Although some elements of the current fiscal response (particularly in the US) offer an alternative approach to addressing structural disinflation, in our view, much of that remains contingent on successful passage through the Senate.
    At the same time, there is no shortage of liquidity trying to find yield (credit spreads are little changed more recently, despite rising yields). While we expect yields to rise further this year, the upshot too for bond investors is that the steeper yield curve now offers a more attractive “carry and roll” option. More pointedly, for those with a ‘buy-and-hold’ investment strategy (who are therefore less sensitive to price fluctuation due to their unrealised nature), we see maturities around ten years presenting the optimum balance between outright yield and risk, at present.
  • A laddered portfolio can be useful in mitigating rising yield curves. The idea is to diversify and spread the risk along the interest rate curve to hedge against any idiosyncratic moves in rates. We note that the average duration of client portfolios is less than two years. In aggregate, clients presently have very little exposure to interest rate risk.
  • Consider inflation-linked bonds to protect against inflation, either as a perfect hedge or for those believing inflation will rise higher than currently priced-in. Over the next year or so it is almost certain we will see episodes of rising prices that will be reflected in a higher consumer price index. However, be mindful to the risk of underperformance. If actual inflation is less than the market expects, the bond will underperform a nominal bond (and vice versa). It is also important to highlight that inflation-linked bond investors are either seeking a perfect inflation hedge (e.g. life insurance companies) or betting that the market’s inflation expectations underestimate the investors’ own view on future inflation. The first category will have very little sensitivity to actual inflation, while the second category will be less focused on headline inflation and more on the difference between expected and actual inflation. We also note that the AUD inflation-linked corporate bond market is highly illiquid, with prices more often fluctuating due to supply and demand rather than market or credit fundamentals which then introduces additional complexity for this instrument type.