COVID-19 cases and hospitalizations are in retreat across much of the world (see next chart). Even Japanese infections – which had recently surged to record levels – are now in precipitous decline (see subsequent chart).
COVID-19 cases are in retreat
As of 09/12/22. Change in cases measured as the 7-day change of the 7-day moving average of daily new infections. Source: World Health Organization, Macrobond, RBC GAM
COVID-19 cases and deaths in Japan are falling after record levels
As of 09/11/22. 7-day moving average of daily new cases and deaths. Source: Our World in Data, Macrobond, RBC GAM
Happily, two Omicron-targeting vaccines have been approved or are in the process of being approved, depending on the jurisdiction. These should be significantly rolled out over the coming months, providing additional protection and hopefully preventing any winter wave from inflicting too much damage.
Further, it is promising that no new variants have yet emerged to displace the currently dominant BA.5 Omicron sub-variant. It is still prudent to assume there will be future waves driven by new, more contagious variants (or even by waning immunity to BA.5). However, the waves did eventually end for the Spanish Flu.
China remains the exception to this positive COVID-19 narrative. Its zero-tolerance policy has most recently ensnared Chengdu, a city of 21-million people. The city has been put on indefinite lockdown. In fact, according to CNN, more than 70 Chinese cities have been put under a full or partial lockdown since late August, affecting more than 300 million people. This may be a particularly vigilant push in the lead up to the National Congress that takes place on October 16. It is plausible that restrictions could ease thereafter, but that is mere speculation.
Ever since Ukraine rebuffed a portion of Russia’s initial assault in late February, Russian forces have been nibbling away at Ukrainian territory. That trend appeared to reverse over the past week, as Ukraine reclaimed a chunk of its land in the east of the country, near the city of Kharkiv.
It still seems reasonable to expect Ukraine to gradually achieve the upper hand given the ongoing supply of western weaponry versus the gradual diminishment of Russian stocks. However, this is far from certain and the war is likely to persist for quite some time. Sanctions should endure for even longer – the observation of greatest relevance from an economic and inflation standpoint.
Russia is now effectively sanctioning itself with regard to the sale of natural gas, having steadily reduced its provision of the commodity to Europe, and cutting it off altogether in early September. Natural gas prices surged another 30% on this news. At current prices, the European Union is on track to spend seven times more on electricity and gas than normal over the coming year.
Reflecting the altered energy landscape – the diminished reliability of Russian supplies combined with the mounting importance of securing domestic energy security – Japan has announced it will return to nuclear power after shunning the technology over the eleven years since the Fukushima nuclear disaster. The government aims to revive 17 out of its 33 operable nuclear reactors by next summer.
We begin with a refresher of key inflation views, before proceeding in several granular directions.
North American inflation appears to have peaked in June, softened in July and should be muted again when the August data becomes available. While it is unlikely that monthly price changes will remain quite that limp, we nevertheless budget for generally diminishing inflation pressures ahead, based on the view that all four of the key drivers of high inflation have turned:
- Monetary stimulus has become monetary restraint.
- Fiscal stimulus has similarly become a fiscal drag despite recent new initiatives.
- Supply chain problems are significantly resolving (more on that later).
- The commodity shock – ex-natural gas – has materially unwound.
The most likely scenario is that inflation diminishes over the coming six months, approaching – if not quite fully reaching – a historically normal monthly rate of change. But this forecast, as reasonable as it is, is not the only conceivable scenario (see next graphic).
U.S. inflation scenarios suggest return to normal
As at 09/08/22. Source: RBC GAM
We assign a 55% chance to this base-case scenario, alongside a 20% chance that inflation remains too high. Technically, this is split into a 10% chance that inflation remains at current levels and a 10% chance that it ascends further. There is also a 25% chance that inflation falls below normal.
The existence of a low inflation scenario might be surprising – particularly the notion that it is more likely than the scenario in which inflation remains too high. But with the commodity shock and supply chain problems unwinding, there could be not just a retreat in price pressures but an outright reversal as the underlying cost of shipping products and procuring goods reverses some of their earlier advance. Recessions usually bring powerful disinflationary forces as well.
A key goal in this scenario-building exercise is to explicitly acknowledge the low inflation scenario, as it is underappreciated.
As depicted in the graphic above, we also contemplate the medium-term inflation outlook. This should be interpreted as representing inflation in one to two years’ time. We believe there is an 80% chance that inflation eventually settles at a broadly normal level (though allowing for the definition of normal to include inflation up to about 3%). That leaves a 10% chance that inflation remains enduringly too high, and a 10% chance that inflation remains enduringly too low. The former scenario represents the failure of central banks or the arrival of additional inflationary shocks. The latter scenario would reflect a deeper recession and/or a stronger tendency for distorted prices to retreat back to their prior levels.
European gas subsidies
European governments are recognizing that households, businesses and certain power companies cannot be forced to fully absorb skyrocketing natural gas and electricity prices. A range of measures have thus been implemented to shift the burden from the private sector to the public purse:
- Germany has now implemented three energy support packages, collectively costing 95 billion euros.
- The French relief package is expected to cost 64 billion euros.
- Italy is thought to have already spent 52 billion euros.
- The U.K., under new Prime Minister Liz Truss, has promised to freeze energy costs for two years, to the tune of 100 billion pounds.
These actions have several important consequences.
- They limit how high consumer prices will rise, since those prices are now being artificially capped in several countries. As such, for instance, the U.K. annual Consumer Price Index (CPI) no longer appears likely to reach the high teens, though Europe and the U.K. will still likely suffer through higher inflation than North America in the near term.
- Some governments are looking into placing temporary windfall taxes on energy companies that are profiting from high energy costs. This would discourage investment in future capacity, but help to pay for the energy subsidies.
- Fiscal deficits will be significantly larger than before due to these initiatives, at a time when quantitative-easing operations were already coming off. That would suggest, all else equal, higher yields.
- Direct subsidies are a flawed solution in that they don’t incent people to use less energy. And without that, the energy shortage could be quite serious indeed over the winter. A better solution would be rebates unconnected to actual energy usage, or capped at a limited level of energy usage, that reward people for cutting back.
- More support is likely in future years. To the extent it may take several winters for Europe to fully shift its energy consumption away from Russia and natural gas, there will be pressure for these programs to be extended for years. The U.K. is already promising support for two winters. This adds to the cost.
- At least so far, most European countries remain firm in supporting sanctions despite Russian pressure. An exception is Bulgaria, which is thought likely to strike a deal with Russia.
The shelter component of inflation continues to run hot, even as housing markets begin to cool. This is not a new phenomenon: the shelter part of the Consumer Price Index (CPI) is famously lagged relative to the housing market.
A recent White House report estimates that it takes 16 months for a turn in home prices to be picked up in the owner’s equivalent rent portion of shelter costs in CPI. To the extent that the U.S. Case-Shiller home price index hasn’t even begun to really turn, this suggests that shelter will remain a source of heat for some time to come, though it should eventually cool. When it does start to cool, it should remain a disinflationary force for an extended period of time.
In Canada’s case, the lag should be somewhat less. Not only have Canadian home prices already been falling for six months, but dwelling costs are incorporated into Canadian CPI in a way that should render them somewhat less lagged. Still, a substantial lag remains: the owned accommodation portion of Canadian shelter costs has only decelerated slightly so far.
U.S. regional variations in inflation
Inflation is not the same everywhere within a country. In the case of the U.S., it is interesting – and logical – to note that the lowest inflation has been in the Northeast, whereas the highest inflation has been in the South. Loosely, this aligns with where Americans have moved from and to during the pandemic. With a rising population home prices rise by more, the economy runs hotter and products are in shorter supply. All of this is inflationary for the South.
There may also be a non-pandemic structural element at play since the Northeast has been losing population share to the South and the West for many years.
Northeast inflation is running 1.2 percentage points below the U.S. average. South inflation is 0.9 percentage point higher. The West and Midwest land in the middle.
Shrinkflation is a sneaky kind of inflation in which the size of a product subtly shrinks rather than the price of the product rising. Naturally, it is more common during times of high inflation as companies become desperate for ways to pass along rising input costs.
Recent examples include the size of a Gatorade bottle shrinking from 32 ounces to 28 ounces (equating to a 14% price increase) and Domino’s Pizza reducing its chicken wing order from 10 pieces to 8 pieces (a 25% price increase). These changes happen most frequently in the food and drink space.
Shrinkflation is undoubtedly sneaky and means that inflation is higher than it appears to be for the average person. However, contrary to what one might imagine, it is already properly captured in consumer price indices (CPI). CPI is already measured on a per unit basis where possible, so this extra inflation does appear in the official estimates.
Under normal circumstances, the magnitude of the shrinkflation effect on inflation is small. A pre-pandemic U.K. study could only find a significant effect for confectionary items and identified a mere 1.2 percentage point cumulative effect on the sector’s prices over a five-year span. Mapped onto the overall consumer price index, shrinkflation contributed just 0.03 percentage points to inflation over the period. One imagines the effect is several times that today, though the resultant sum would still be fairly small.
Complicating matters, sometimes a company reduces the size of its product but claims that rising quality fully offsets the change. For example, Folgers coffee reduced its standard container from 51 ounces to 44 ounces, but insists that the quality increased proportionally, such that 400 cups of coffee can still be brewed from the container. This deflationary effect would not be captured in CPI.
The lesser-known service sector equivalent of shrinkflation is called skimpflation. An example is that, post-pandemic, hotel rooms no longer receive housekeeping service every day. One might also argue that the provision of online education and online health care during the pandemic was of a lower quality than the in-person equivalents before the pandemic. But the cost did not generally decline, meaning there was extra inflation. Unlike with shrinkflation, statistical agencies do not have a mechanism for adjusting their inflation estimates to account for skimpflation. As such, some service sector components likely have inflation rates somewhat higher than officially registered.
Supply chain inflation
Inflation deriving from supply chains should reverse as supply chains themselves heal. Examples of important improvements include plummeting container shipping costs (see next chart), a return to nearly normal dry bulk shipping costs (see subsequent chart) and a sharp decline in the anxiety of manufacturing purchasing managers about both inflation and their supply chains (see third chart).
Shipping costs fall further
As of the week ending 09/01/22. Source: Drewry Supply Chain Advisors, RBC GAM
Shipping costs retreat from latest peak
As of 09/09/22. Shaded area represents recession. Source: Baltic Exchange, Macrobond, RBC GAM
Price increases and supplier deliveries returning to normal ranges
As of July 2022. Shaded area represents recession. Source: Institute for Supply Management, Macrobond, RBC GAM
Second-round inflation pressures
As the primary drivers of inflation turn, a key remaining question is the extent to which second-round inflation pressures will pick up their mantle and keep inflation roaring. Three key considerations are the breadth of inflation, wage pressures and inflation expectations.
- With regard to inflation breadth – the extent to which a lot of different products are becoming more expensive – the news is still grim. The breadth of high inflation again increased (slightly) in July. Inflation was softer in July for a very narrow reason: the decline in gasoline prices. The breadth of high inflation still risks high inflation becoming self-perpetuating.
Inflation in the U.S. has broadened significantly
As of July 2022. Share of CPI components with year-over-year % change falling within the ranges specified. Source: Haver Analytics, RBC GAM
- Wage pressures may be starting to turn downward. Overall U.S. wage growth has ebbed only slightly, but the bellwether limited-service restaurant sector has experienced a sharp deceleration in wages (see next chart). This sector includes fast food restaurants, which employ lower skilled workers who are the last to be hired and the first to lose their bargaining power as labour markets sour.
Wage growth of U.S. low-skilled workers decelerating
Limited-service restaurants as of July 2022, total private nonfarm as of August 2022. Source: Bureau of Labor Statistics, RBC GAM
- Inflation expectations continue to ease. While financial market-based inflation expectations declined some time ago, main street inflation expectations were more reluctant to drop. For inflation to sustainably decline, businesses and households need to be convinced as well. Fortunately, there has recently been a palpable if small drop in business inflation expectations and a slight decline in consumer inflation expectations (see next chart).
U.S. inflation expectations have started to fall
Market-based expectations as of 09/02/22, survey-based consumer and business expectations as of August 2022. Source: Federal Reserve Bank of Atlanta, Federal Reserve Board, University of Michigan Surveys of Consumer, Haver Analytics, RBC GAM
Further emphasizing that businesses are starting to think differently about the inflation outlook, the fraction of U.S. businesses looking to raise prices has begun to descend from record levels (see next chart).
Fraction of U.S. businesses planning to raise prices still high but falling
As of July 2022. Shaded area represents recession. Source: National Federation of Independent Business (NFIB) Small Business Economic Survey, Macrobond, RBC GAM
Economic developments vary widely
Recent U.S. economic data has not been bad:
- The Institute for Supply Management (ISM) Manufacturing index for August held steady at a so-so 52.8.
- The ISM Services measure rose from 56.7 to a solid 56.9.
- U.S. payrolls for August added a further 315,000 jobs, dampened only slightly by a -107,000 negative revision to prior months and a small increase in the unemployment rate from 3.5% to 3.7%.
- Initial jobless claims have also fallen for the last three weeks, albeit after a longer and larger increase over the prior few months.
- U.S. Q3 Gross Domestic Product (GDP) is now tracking a 2-3% annualized gain, suggesting a recession is not yet underway.
The story in Canada, in contrast, is considerably worse. For instance:
- The Canadian job market shed workers for a third consecutive month (see next chart), losing 39,000 overall positions and 77,000 full-time ones.
- The unemployment rate accordingly rose from 4.9% to 5.4%.
- Although there is some suspicion about the seasonal factors affecting certain sectors – it is hard to fathom that 50,000 Canadian education workers actually lost their jobs in August – there is a tinge of truth in other areas. Construction employment fell by 28,000 at a time that the Canadian housing market is weakening.
Canadian labour market shows sign of weakness
As of August 2022. Source: Statistics Canada, Macrobond, RBC GAM
The Canadian economy has also lately been stumbling. After recording a month of flat economic output in May, June notched a muted 0.1% gain and July is tracking -0.1%. That is effectively a quarter without economic growth.
Soft versus hard data
“Soft” economic data in the U.S. – surveys and the like – have weakened quite a lot, whereas “hard” economic data – measures of actual spending and hiring – have not descended to the same extent (see next chart).
Both hard and soft economic soft data in U.S. turn negative
As of 09/01/22. Source: Citigroup, Bloomberg, RBC GAM
The most likely scenario is that the hard data eventually follows the soft data, fulfilling the recession prophecy. Indeed, it is logical to think that sentiment and expectations should drop before actual activity declines. But one cannot completely rule out the “soft landing” scenario in which weak sentiment never fully translates into poor spending and hiring. The pandemic era of economic activity has been sufficiently topsy-turvy that weird disconnects remain possible.
Consensus outlook continues to weaken
It is unsurprising but still instructive to note that the consensus economic outlook for the U.S. economy – and for most economies – continues to weaken. In the case of the U.S., the decline has accelerated of late, with the 2022 and 2023 consensus GDP estimate in near free-fall (see next chart). Our own forecasts remain below the consensus. It is also notable that the consensus expectation for 2022 is weaker than for 2021, and that the 2023 outlook is weaker than for 2022.
U.S. consensus growth forecast continues to decline
As of July 2022. Source: Consensus Economics, International Monetary Fund, RBC GAM
Labour market debate
There remains a fierce debate about what to expect from the labour market as economic activity softens. Much of this revolves around the contours of the Beveridge Curve, which examines the relationship between job openings and unemployment (see next chart).
U.S. Beveridge curve – job openings vs. unemployment rate
As of June 2022. Source: BLS, Job Openings and Labor Turnover Survey (JOLTS), Macrobond, RBC GAM
Optimists note that the level of job openings is significantly higher than normal for the present level of unemployment. This presumably reflects worker-employer matching problems in the economy given the many distortions wrought by the pandemic. To the extent pandemic distortions fade and/or time passes, these mismatches should gradually resolve. The hope is that as the economy weakens, companies will shelve their rather extensive hiring plans rather than lay off existing workers. This would be unusual but plausible, in part because it would merely pull the Beveridge Curve down to its historically normal stomping ground.
Pessimists, conversely, note that there is no precedent for job openings falling without there also being a significant loss of actual jobs. They argue that the natural unemployment rate appears to have increased, and so the economy must eventually capitulate to that reality. Reducing the overheating in the economy likely requires a multi-percentage point increase in the unemployment rate. Furthermore, there has been a recession in the U.S. any time the unemployment rate rises by even a small amount.
Both sides make good points. The true answer likely lands somewhere in the middle, though our thinking is skewed somewhat toward the pessimists’ side of the equation. This is to say, the unemployment rate probably won’t rise as much as normal, but cooling the economy likely requires a recession and will result in significant job losses.
Canadian housing weakens
Under normal circumstances, the Canadian housing market arguably generates more than 20% of Canadian economic output when one combines the output produced by residential construction, renovations, realtors, real estate lawyers, mortgage lenders, the purchase of appliances and furnishings, plus the wealth effects normally unleashed by rising home prices.
Today, Canadian home prices are most certainly not rising, but there is some confusion as to precisely what home prices have or haven’t done recently.
Naturally, home prices have fallen more sharply in jurisdictions that experienced the greatest run-up in prices beforehand. There is considerable regional variation.
There are also wildly varying estimates at the national level. The most exciting claim is that nationwide home prices have fallen by 17%, but this is not an accurate estimate as it comes from an unweighted MLS metric that fails to control for the shifting composition of home sales. If more luxury homes are sold one year and more condos the next, it appears that home prices have collapsed, whereas the individual price of luxury homes and condos may not have changed at all.
At the opposite extreme are the Teranet-National Bank Home Price Index and the New Housing Price Index. The former argues that home prices have barely fallen (-0.3%), while the latter asserts that they are still actively rising. But these measures also have significant flaws.
The Teranet measure only examines paired home sales – transactions for which a particular house has been sold twice in their databank. This greatly limits the sample size. Historically, the Teranet measure has significantly lagged the turning points captured by more timely measures – a significant knock against it if we want to track home price movements in a timely way.
For its part, the New Housing Price Index completely excludes a major segment of the housing market (condos), the land value portion of the index is obtained quite loosely (estimated by homebuilders), and – crucially – new homes are usually built on the outskirts of an urban area, with each generation of new homes built further away from downtown and thus less valuable. It is not a proper like-for-like comparison. Empirically, the index is also extremely smooth, rarely identifying big upswings or downdrafts (see next chart).
Housing market correction underway in Canada
As of July 2022. MLS HPI and Teranet-National Bank HPI measure changes over time in the resale prices of existing homes. Source: Canadian Real Estate Association, Teranet/National Bank of Canada, Statistics Canada, Macrobond, RBC GAM
Our favourite measure lands in the middle of these extreme estimates. It is also from MLS, but calculates a benchmark price by properly controlling for the type and size of dwelling sold. It estimates a 9.9% decline in home prices so far (see next chart).
Run-up in home prices dwarfs the decline so far
As of July 2022. MLS HPI and Teranet-National Bank HPI measure changes over time in the resale prices of existing homes. Source: Canadian Real Estate Association, Teranet/National Bank of Canada, Statistics Canada, Macrobond, RBC GAM
We downgraded our Canadian home price forecast a few months ago, and are now looking for a cumulative 20-25% peak to trough decline. That would take home prices most of the way – but not quite all of the way – back to pre-pandemic levels. So far, 30% of the price increase has been unwound.
U.S. student loan forgiveness?
The Biden administration announced a large executive order several weeks ago that proposes to spend in the realm of $500 billion to cut $10,000 to $20,000 from federal student loan balances for those with less than $125,000 in income. The order also delays student loan interest payments until 2023.
This is a significant fiscal outlay, albeit one that is spread over many years since the cancellation of a loan results in the loss of a stream of interest and principal payments spanning many years.
In the gripes department, this does little to address the problem of people with huge student debts. It punishes those who sacrificed to pay off their student loans more quickly. It represents a cash transfer to the most educated segment of society, and it doesn’t apply to future graduates. It also misses the most enduring solution, which would be to reduce the eye-watering and rapidly inflating cost of a university education in the U.S. But an executive order is not legislation, and so is limited in what it can accomplish. Legislation is unlikely given insufficient votes in the Senate.
But we have buried the lede: there are serious doubts that the White House can actually deliver on this promise. Significant legal challenges are expected, and the executive order is thought by some legal scholars to be an overreach. It is a political win for the White House, either way: they either get to implement the policy or if it is blocked they can rally voters for the midterm election.
One thing that will prove consequential regardless of the result of legal challenges: those with student loans have not had to pay interest on their loans for the duration of the pandemic. Those payments will recommence by January. With US$1.6 trillion in student loans and, quite conservatively, a 5% interest rate, that represents US$80 billion in additional interest payments that will subtract from consumer spending and could cause some households financial distress at a time of rising rates and high inflation.
Recession risk remains high
We continue to assign a 70% chance of a U.S. or Canadian recession by the end of 2023, with an even higher risk in Europe and the U.K. It is worth highlighting a few recent inflation signals.
The U.S. 2-year to 10-year spread has long been inverted, and now the inflation-adjusted spread has also inverted (see next chart). It is not classically used to gauge recessions, but arguably carries more informational value than usual right now. It helps to strip away distortions that might arise from the expectation embedded within nominal yields that inflation eases over time.
Real yield curve inverted after reaching multi-year high
As of 08/22/22. Source: Bloomberg, RBC GAM
Meanwhile, another classic recession signal – the slope of the 3-month yield to the 10-year yield – is now very close to inversion. In fact, it is probably just a matter of weeks as the Federal Reserve raises rates, which ratchets the 3-month yield higher over time. Whereas the 2-10 curve inversion tends to lead the recession by 18 months and the peak in the S&P 500 by 15 months, the 3m-10 curve inversion tends to happen closer to those events (11 months and 10 months ahead, respectively). It should be conceded that, in our own minds, a recession is likely to happen sooner than that and the S&P 500 peak has probably already passed.
Even though jobless claims have recently edged back downward in recent weeks, it caught our attention that the rise in jobless claims from trough to recent peak was – barely – enough to trigger another historical recession signal (see next chart). The fact that jobless claims have since retreated doesn’t invalidate the signal, as per the experience in the early 1990s and the early 2000s. One curious thing is that, historically, the rise in jobless claims happened well into the recession, whereas this time the recession is still merely anticipated.
Rising U.S. initial unemployment claims are signalling recession
As of week of August 15 2022 (week 33). Source: U.S. Department of Labor, Macrobond, RBC GAM
On a related note, we stick with the view that a Canadian recession should be worse than in the U.S. Indeed, we recently ran an interest rate hike scenario through our large-scale econometric model and found that, as expected, the Canadian economy and housing market are more rate-sensitive than in the U.S.
Monetary tightening continues
Aggressive monetary tightening continues. The Bank of Canada raised rates by another 75 basis points; the European Central Bank hiked by 75 basis points; the Reserve Bank of Australia raised rates by 50 basis points; and the U.S. Federal Reserve is expected to raise the fed funds rate by 75 basis points on September 21.
The Canadian rate decision highlighted two important themes.
- The Bank of Canada refused to celebrate despite weaker inflation in July. The weakness was identified as being the mere result of lower gas prices, with inflation breadth described as continuing to expand. This is factual, though one could simultaneously argue that August inflation is likely to be muted and that certain key inflation drivers are turning. But, for central banks, the importance of reclaiming their inflation-fighting credentials is such that they cannot afford to flag or claim even a smidgen of success until total victory over inflation is assured. As such, central banks remain likely to err on the side of tightening too much rather than too little.
- The Bank of Canada has certainly not reached peak interest rates, but it has probably passed the point of peak tightening. The first derivative has turned. The central bank tightened by 100 basis points in July and by 75 basis points in September. More muted language about future actions has convinced markets to price a 50 basis point rate increase for October, followed by a 25 basis point hike for December. This is a steady deceleration, conceivably landing at a 4.00% peak policy rate to ring in the New Year.
Markets expect a similar peak policy rate of about 4% in the U.S. (see next chart).
Expected peak fed funds rate has increased sharply
As of 08/30/22. Source: Bloomberg, RBC GAM
-With contributions from Vivien Lee, Vanessa Adams and Aaron Ma