Yields rise even further
Even though several developed-world central banks have tentatively stopped raising their policy rates, bond yields have continued to rise. The U.S. 10-year yield has increased from as little as 3.28% in the spring to a heady 5.00% in recent weeks (see next chart). That represents the most elevated bond yield in 16 years.
U.S. yields rise on resilient economic data, Fed’s higher for longer stance and supply influx
As of 10/26/2023. Shaded area represents recession. Sources: U.S. Treasury, Macrobond, RBC GAM
The increase in yields does not reflect more hawkish expectations for central banks, nor does it convey rising inflation expectations. Instead, it is mostly the result of a higher term premium (see next chart). At essence, a number of factors have altered the supply-demand dynamic for sovereign bonds, such that higher interest rates are necessary to pull enough bond buyers into the market to absorb the current level of issuance.
Term premium has climbed sharply
As of 10/25/2023. Sources: Federal Reserve Bank of New York, Macrobond, RBC GAM
Drivers of a bigger term premium in the U.S. include:
elevated bond issuance to fund the large fiscal deficit,
the steady sale of bonds held by the Federal Reserve,
the recent U.S. debt downgrade (which argues for a larger risk premium embedded within the term premium),
greater competition for capital with Japan and Europe as those regions emerge from negative interest rates,
China’s sales of some of its foreign bond holdings as the country defends its currency.
Although the term premium has increased greatly, it cannot be said to be unreasonably high. Instead, term premiums were unnaturally compressed – indeed, outright negative – for the bulk of the last decade. That was not the normal state of affairs. It was distorted by quantitative easing, negative interest rates in other countries and a global savings glut.
From an economic standpoint, higher interest rates argue for even greater economic weakness. They also incrementally raise the likelihood of a recession.
While we anticipate lower bond yields in a year’s time, this is mainly because we think central banks will eventually be in a position to cut rates closer to neutral levels. This will pull longer-term yields downward without collapsing the newly revived term premium.
From an economic standpoint, higher interest rates argue for even greater economic weakness. They also incrementally raise the likelihood of a recession (discussed later).
True, the U.S. economy is less interest-rate sensitive than it once was. Households have deleveraged since the global financial crisis. Thirty-year mortgage terms have grown in popularity. And Moody’s estimates that just 11% of household debt now fluctuates in real time with interest rates. That figure was more than twice as high during the global financial crisis, and nearly four times higher during the last monetary policy-driven recession of the late 1980s/early 1990s.
Those on the optimistic side of the debate may be misattributing the economy’s resilience to reduced rate sensitivity when the bigger reason is that it takes a long time for higher interest rates to fully impact the economy.
But U.S. rate sensitivity is not nil. Households are already paying more interest as a share of disposable income than at any point in the last decade and the adjustment is merely beginning (see next chart). Furthermore, interest rate sensitivity isn’t just about what fraction of loans have a variable rate. It’s also about what the economy is accustomed to.
Borrowing costs were stunningly low for the bulk of the 2010s, making the increase over the past two years even more startling and consequential. In contrast, in the monetary tightening cycle of the late 1980s, interest rates were rising but remained far lower than the norm of the prior decade – and yet a recession still took hold. The risks are thus considerable today.
U.S. household debt service costs are rising
As of June 2023. Sources: Federal Reserve, U.S. Bureau of Economic Analysis (BEA), Macrobond, RBC GAM
Those on the optimistic side of the debate may be misattributing the economy’s resilience to reduced rate sensitivity when the bigger reason is that it takes a long time for higher interest rates to fully impact the economy. The lag in the U.S. from a first rate hike to recession has historically averaged 27 months. Had a recession already arrived, it would have been unusually early. The window remains wide open for a monetary policy-induced downturn.
Providing some international perspective, not only have a number of global central bankers posited that the lag from higher interest rates to the economy is longer than normal this cycle, but a recent survey of Canadian businesses found that the largest fraction believes the impact of higher interest rates is “just beginning.”
Even with a greater orientation toward fixed-rate lending, borrowers roll into higher rates faster than one might think:
While the most popular U.S. mortgage term is a 30-year fixed rate, the median duration of home ownership is just 13 years. This means that something like 8% of existing homeowners will be exposed to a higher mortgage rate each year, despite theoretical protection for much longer.
The average auto loan is for 72 months. This means that around one-sixth of car loans become exposed to higher interest rates each year – and a new car loan will set you back around 10% in annual interest right now.
The average term bank loan to businesses lasts 5—7 years, meaning that around 15—20% of these are rolling into higher interest rates each year.
Credit cards are effectively a variable rate product, with an immediate impact from higher interest rates. The U.S. 30-day credit card delinquency rate is already rising (see next chart).
Credit card delinquency rates surge
As of Q2 2023. Shaded area represents recession. Sources: Federal Reserve Board, Macrobond, RBC GAM
Critically, it is important to understand that for every borrower with a lovely locked in mortgage at low rates, there is also a lender locked in on the other side of that low rate, and suffering for it. It is unlikely a coincidence that the U.S. with its locked-in mortgages and longer-term bonds has had the greatest trouble with its financial institutions recently. These carry substantial bond market losses (much unrealized, as per the next chart), never mind losses on their loan portfolios.
In response, U.S. banks have tightened their lending standards more fiercely than in most countries and are in a position to lend less. As a result, the U.S. may be less exposed to the price effect of rising interest rates, but it is consequentially more exposed to adverse quantity effects.
Unrealized losses rise on investment securities held by FDIC-insured institutions
As of Q2 2023. Sources: Federal Deposit Insurance Corporation (FDIC), Macrobond, RBC GAM
Finally, in discussing the economic effect of higher interest rates, it is important to recognize that economic output is fundamentally generated by the production of new things: new homes, new cars, new factories, and so on. Anyone buying a new home, car or factory is fully exposed to the prevailing interest rate, and so must decide whether or not to proceed on that basis. The U.S. economy enjoys no greater protection in this capacity than any other. Decisions to scale back on such activity will be no less damaging there.
The point of this diatribe is that higher interest rates will likely prove economically painful, even for the U.S.
Yield curve un-inverts
In an inverted yield curve, longer-dated bond yields are lower than shorter-dated bond yields. When the inversion persists, it has historically signaled a recession to come, and so has constituted a key argument for an approaching recession.
But the yield curve has become significantly less inverted in recent months (see next chart). The U.S. 2-10 year spread was as negative as -100 basis points, whereas now it is just -36 basis points. The reversal along other portions of the yield curve has been even greater.
Yield curves un-invert lately on expectations of higher rates for longer
As of 10/24/2023. Near-term forward spread measured as forward rate of 3-month Treasury bill six quarters from now minus spot 3-month Treasury yield. Shaded area represents recession. Sources: Engstrom and Sharpe (2018), FEDS Notes, Washington: Board of Governors of the Federal Reserve System, Bloomberg, Haver Analytics: RBC GAM
How should this be interpreted with regard to the risk of a pending recession?
A first observation is that the yield curve often steepens back up immediately before a recession. An inverting curve is a good gauge of a recession approaching from perhaps a year out, while aggressive steepening (and thus un-inverting) often happens right before the recession as central banks snap into rate cutting mode, pulling short-term yields downward.
But this observation is somewhat of a red herring in the present situation. The recent steepening has been a bear steepener rather than the usual bull steepener. In other words, it has happened because long-term interest rates have been rising rather than short-term interest rates falling. This does not normally precede a recession.
So has the risk of a recession simply diminished? We are inclined to argue not.
The increase in longer-term rates did not arise due to more hawkish central bank expectations, from an improving economic outlook or from rising inflation expectations – all of which might reflect a lower recession likelihood. Instead, technical supply and demand factors have driven the term premium higher.
Without a direct recession message, all you can really say is that the increase in longer-term rates is itself a drag on the economy and so should theoretically increase the risk of recession as a second-order implication.
Incidentally, we had warned early in the curve inversion that the recession implication wasn’t as overwhelming as it looked if you adjusted for the non-existent term premium. Now, with a rejuvenated term premium, the curve is only moderately inverted. But this actually makes it more akin to past recessions than before.
Recession probability rises
Swirling forces continue to impact the probability for a recession. Using the U.S. as our benchmark and forecasting out over the next 12 months, we have upgraded the probability of a recession from 65% to 70%.
Arguments in favour of this increase include:
Interest rates have increased significantly, with a theoretically detrimental effect on growth.
Inflation is running a little hotter than previously expected, providing less scope for rate cuts (and highlighting the risk of another rate increase).
The geopolitical environment has become more complicated. While the most likely scenario does not materially alter the likelihood of recession, there are downside scenarios that do.
Mounting economic weakness outside of the U.S. further highlights the negative impact of higher interest rates. The interconnected nature of the global economy argues that no country is likely to be completely immune should others stumble.
Unemployment rates are clearly no longer falling and have tentatively – if sporadically – begun to rise. Historically, this has been a strong recession signal.
We continue to collect evidence that higher interest rates take a surprisingly long time to fully impact economic activity.
Of course, it is not a one-way street. A few arguments against a recession have strengthened:
Initial jobless claims have fully unwound an earlier increase that had temporarily triggered a recession signal.
The U.S. economy, in particular, has so far refused to quit.
The yield curve is less inverted than before (though, as discussed, we are dubious that this truly means the risk of recession has fallen).
On balance, we see more negatives than positives and so have increased the assigned recession likelihood. But this won’t be the last word on the subject.
Markets signal economic weakness
We are normally tasked with using economic signals to predict financial markets. However, financial markets can also be used as a leading indicator for the economy. We tend to rely on them as a helpful cross-check of our economic views.
In that context, the stock market is pointing to economic weakness ahead. While the S&P 500 is approximately 7% higher than it was at the beginning of the year, this greatly overstates the market’s enthusiasm about the prospects for the average business. The “Magnificent Seven” tech stocks are a remarkable 56% higher over 2023, whereas an equal-weighted S&P 500 index is actually down 5% (see next chart). That better conveys the average experience.
U.S. equity indices show signs of weakness
As of 10/27/2023. Magnificent 7 is a cap-weighted index which includes Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta and Tesla. Sources: Bloomberg, RBC GAM
Stock market conditions have then gone from simply lacking breadth to demonstrating outright weakness over the past three months. The S&P 500 is down by 9% since the start of August, putting it on the cusp of a formal correction. Internationally, the MSCI World index is already there, down 11% from its August peak.
For their part, credit markets have expressed less concern, but spreads there have also increased slightly in recent weeks (see next chart).
U.S. credit spreads have narrowed, but have crept higher lately
As of 10/26/2023. Spreads of ICE BofA U.S. Corporate Index. Shaded area represents recession. Sources: International Currency Exchange (ICE), Bloomberg
Financial markets are expressing more nervousness about the macro environment than they have since the fall of 2022. In fairness, the economy held up better than expected at that time and so the same trick could well be achieved a second time. But it is arguably unwise to repeatedly bet on upside surprises. For our part, we continue to believe a recession is more likely than not over the next year.
This is the first #MacroMemo published since the onset of the war between Hamas and Israel on October 7. At this point, the tragic details of the conflict to date have been well documented and so we won’t reiterate them here.
Turning, instead, to an economic perspective, the implications are fairly small in the short run, at least outside of the directly affected areas. Global risk aversion briefly rose when the war began, as demonstrated by a stronger U.S. dollar and lower bond yields. But these responses quickly dissipated.
There is undeniably an upside risk to the price of oil if this conflict were to unexpectedly broaden. Oil stocks are approximately 8% higher than they were before the conflict began, flagging that the risk is not nil.
The price of oil temporarily rose by around US$5 per barrel, a classic response whenever geopolitical tensions rise in the Middle East. But this has also since abated, and of course neither Israel nor Gaza produce any oil of note.
Geopolitical developments in the Middle East normally play out in financial markets via their effect on oil prices. The reversal of the oil market’s initial knee-jerk leap conveys the expectation that major oil powers such as Iran will not ultimately become involved. Indeed, OPEC nations as a collective do not appear to be in the mood for the sort of embargo that ignited oil and inflation in the 1970s.
Incidentally, the oil intensity of the global economy is 56% less than it was in 1973, making any theoretical embargo less devastating. Still, there is undeniably an upside risk to the price of oil if this conflict were to unexpectedly broaden. Oil stocks are approximately 8% higher than they were before the conflict began, flagging that the risk is not nil.
In practical terms, the Middle East had been on a gradual stabilizing path for several years prior to this event, prominently including a formal normalization of relations between Israel and the United Arab Emirates in 2020. Economic interests and the understanding that greater stability in the region benefits everyone have been prime motivators. Israel had been on the cusp of signing a similar pact with regional giant Saudi Arabia when the conflict began. That is now, at a minimum, delayed.
Finally, it is possible (though still entirely speculative) that the U.S. 2024 presidential election could be affected by this new conflict. American presidents have famously enjoyed popularity boosts during wars, and while this is not strictly an American war, the U.S. is sufficiently invested in the situation that President Biden’s chance of a second term could increase at a time when polling shows a hypothetical re-match between Biden and Trump to be tightly contested.
Inflation re-accelerated in August and then again in September, driven mainly by higher oil prices that leapt from a low of US$67 per West Texas barrel in June to as much as US$94 in late September – a 40% appreciation. But the price of oil has since retreated to US$84 per barrel (see next chart), and so it appears likely that inflation in October (and hopefully beyond) will be somewhat tamer.
Crude oil prices rose and have now declined
As of 10/26/2023. Sources: Macrobond, RBC GAM
It is promising that real-time inflation metrics are beginning to reflect this anticipated cooling in the inflation rate (see next chart). The Cleveland Fed’s Consumer Price Index (CPI) Nowcast also points to an October inflation rate of just 0.15% month-over-month (MoM), and a 0.34% MoM rate for core CPI – both improvements on the prior two months.
U.S. Daily PriceStats Inflation Index reflect cooling trend
PriceStats Inflation Index as of 10/24/2023, CPI as of September 2023. Sources: State Street Global Market Research, RBC GAM
Admittedly, the acceleration of inflation across the late summer was not solely a function of higher oil prices. Core inflation also heated up. It rose from two consecutive +0.2% MoM readings (which were actually below 0.2% to a second decimal place) to +0.3% MoM readings in August and September.
Accordingly, we are incrementally more worried than before that inflation might get stuck at an elevated level. A model maintained by the St. Louis Fed now acknowledges a heightened probability that inflation remains above 2.5% over the next year (see next chart).
U.S. Price Pressures Measure suggests inflation may be sticky
As of September 2023. The Price Pressures Measure indicates the probability that expected (year-over-year) inflation rate over the next 12 months will exceed 2.5 percent. Shaded area represents recession. Sources: Federal Reserve Bank of St. Louis, Macrobond, RBC GAM
It’s also most useful to examine the 3-month annualized inflation trend. This avoids over-emphasizing ancient news – as in the year-over-year change -- without being too beholden to the quirks of a single month (see next table). This shows that core CPI remains above 3% (+3.1% annualized, to be precise). Median CPI is even higher at +4.0%. And services inflation ex-shelter is still at a lofty +4.7%.
A great deal of progress has occurred from the days of 9% inflation, but the normalization is not done yet.
As of September 2023. Sources: U.S. Bureau of Economic Analysis, Federal Reserve Bank of Cleveland, Federal Reserve Bank of Dallas, Macrobond, RBC GAM
Fortunately, the housing component of CPI – which was by itself responsible for more than half of the increase in the price basket in September – should be capable of decelerating from here. This is thanks to the combination of lags and mounting evidence that the interest rate-sensitive housing market is again cooling.
In short, our base-case scenario remains for inflation to return to a declining trajectory and even to pleasantly surprise relative to the consensus expectation. However, it is probably not realistic to expect a full return to normality over the next year, and there is a risk that inflation remains considerably too high if a recession does not take hold.
U.S. economic data weakens
Two-thirds of the Big Three auto strikes in the U.S. have been resolved, with workers receiving a 25% wage increase spread over four years. The third will presumably be settled before too long given the precedent of the other two deals and as additional production facilities are idled by the strike. Still, a non-trivial amount of economic damage was done over the past six weeks.
Fortunately, the U.S. economy had a great deal of momentum going into the event. Third-quarter gross domestic product (GDP) was just reported at a fantastic +4.9% annualized. A big number had been anticipated, but its realization is no less impressive. Consumer spending was a particular force.
We are already on the cusp of the next American payrolls report, and it will bear close examining given the upside surprise in the last one, in defiance of what had been a gradual deceleration over most of the prior year. The September edition managed to substantially exceed expectations, with 336,000 new jobs – nearly double the prior trend. Initial jobless claims have remained low since, arguing that the October number should be no worse than fine.
However, little fragments of fragility occasionally reveal themselves:
Job openings and the quits rate have both declined materially from their peak.
Temporary employment continues to decline – historically, a bad sign.
Unemployment is off its low. In contrast to initial jobless claims, continuing jobless claims are again steadily trending higher. The headline hiring number for September was strong, but the less closely watched alternatives were both substantially weaker. The ADP survey argued for a mere 89,000 net new workers, while the household survey claimed just 86,000.
Anecdotally, it is getting harder to find jobs in the U.S. There are also a mounting number of reports of companies offering less money to new hires than they had offered for the same position some time before. The labour market is not as white hot as it once was.
The consensus expectation for October is for about 180,000 new jobs per month – a soft-landing type of number. A significant miss to the downside would start tongues wagging about recession (and we have flagged two consecutive months of sub-150,000 job creation as a potentially important threshold). Conversely, another month well above 200,000 would spark further concerns of an overheating economy and persistent inflation problems.
Finally, there is mounting evidence that the U.S. housing market is beginning to cool again, with evidence ranging from our composite of housing sentiment (see next chart), to a broader agglomeration of real estate indicators (see subsequent chart).
U.S. housing market sentiment remains low
As of September 2023. Housing Market Sentiment Composite is constructed based on survey responses on current and expectations for housing market activities and related drivers. Shaded area represents recession. Sources: National Association of Home Builders, Federal National Mortgage Association. The Conference Board, University of Michigan, Federal Reserve Board of Dallas, Mortgage Bankers Association, Macrobond, RBC GAM
U.S. housing metrics show signs of softening
S&P CoreLogic Case-Shiller Home Price Index as of July 2023; building permits, housing starts and existing home sales as of September 2023; employment and National Association of Home Builders (NAHB) Housing Market Index as of October 2023. Sources: U.S. Bureau of Labor Statistics, Census Bureau, NAHB, National Association of Realtors, S&P, Macrobond, RBC GAM
Internationally, economic data is mostly weaker than in the U.S. (see next chart). We discuss Canada’s wobbling economy next. European Purchasing Managers’ Indices (PMIs) recently came in weak yet again.
Economic growth in Eurozone and U.S. takes a different course
As of 10/27/2023. Sources: Citigroup, Bloomberg, RBC GAM
The Canadian economy is clearly wobbling. Quarterly GDP has fallen in two of the last three quarters (and our tracking suggests only modest growth in the next quarter). The single isolated quarter of growth within this thicket of discontent was fortunately strong enough that one cannot call it a recession (doubly so given a lack of steady job losses). However, the risk is clearly mounting.
Canadian retail sales fell by 0.1% in August and is tracking a merely flat performance for September. Adjusted for inflation, this is a material decline in real terms, which is the basis for GDP. Furthermore, this Canadian consumption is even weaker than it first looks. Given the country’s incredible population growth, one would need nominal spending growth of perhaps 0.4% to 0.6% just for the average person to have acquired the same number of goods as the month before. Canada is nowhere near hitting these numbers, meaning that the average Canadian shopper is already retrenching (see next chart).
Canada’s real retail sales per capita have been declining
As of August 2023. Sources: Statistics Canada, Haver Analytics, RBC GAM
Canadian business expectations are also quite poor:
The Canadian Federation of Independent Business’ barometer just tumbled from a weak 48.8 to a weaker 47.2 – the worst reading since April 2020.
The Bank of Canada’s Business Outlook Survey continues to get worse with every reading, and is now approaching past recession thresholds (see next chart).
It is a similar story with the outlook for future sales (see subsequent chart).
Hiring and capital expenditure plans are also slipping (see third chart).
Inflation concerns are starting to be supplanted by concerns about insufficient demand.
Canadian businesses believe the pain from higher interest rates is only beginning, as highlighted earlier in a different context
Business Outlook Survey Indicator deteriorates further
As of Q3 2023. Sources: Bank of Canada Business Outlook Survey, Macrobond, RBC GAM
Outlook for future sales has been deteriorating
As of Q3 2023. Sources: Bank of Canada Business Outlook Survey, Macrobond, RBC GAM
Canadian businesses pare back on hiring and capex plans
As of Q3 2023. Sources: Bank of Canada Business Outlook Survey, Macrobond, RBC GAM
Finally, while Canada’s labour market has continued to generate new jobs in three out of the past five months, it is unusual that there have been two months of job losses in close proximity. The additional workers don’t seem to be spurring greater economic output given ongoing productivity woes and the latest jobs report managed just 1,000 new private sector jobs out of the 40,000 total positions added.
All of this is to say that one gets the sense of a substantially deteriorating economic environment, and thus the rising probability of a near-term recession in Canada.
China has suffered from a bout of whiplash over the past year. First, the economy shuddered to a halt last fall in response to widespread pandemic lockdowns. Then it lurched forward as the lockdowns were abruptly ended. The economy then stumbled into the summer, before beginning to revive according to the data released over the last six weeks. We have been of the view that there was too much optimism about China at the start of 2023, and too much pessimism recently.
China’s tentative economic rebound is clear in the data: economic surprises have gone from extremely negative to modestly positive in recent months (see next chart).
China’s supporting measures have started to lift economic activities
As of 10/27/2023. Sources: Citigroup, Bloomberg, RBC GAM
A more comprehensive look at key Chinese economic indicators confirms that most are again rising after a period of weakness. We’ve seen especially notable rebounds in exports, property sales and residential construction starts (see next chart).
Monthly economic indicators for China are rising
As of September 2023. Average of 2019 levels indexed to 100. Sources: Haver Analytics, RBC GAM
One wonders about the Chinese stock market, which has continued to hit new lows even as this economic rebound has occurred. This has sent the country’s price-earnings ratio down to a mere 11.5, from nearly 20 three years ago.
A key element of the country’s rebound has been the delivery of a range of stimulus. No single action has been overwhelming, but we’ve seen a mix of actions that in combination have proven sufficient to stabilize the economy, including:
modest rate cuts (including another action in September)
the stimulative effect of a weak renminbi
the partial bail out of beleaguered local governments (including a new announcement of special refinancing bonds by 20 local governments)
tentative Hukou reform.
Hukou is a system of household registration used in mainland China. It assigns individuals as either a rural or urban resident and determines where each type of resident can live, work, and access rights and entitlements such as education, healthcare, and other social services.
Reflecting greater government outlays, the country is targeting a larger budget deficit than before.
Further policy action is possible ahead. Twice a decade, China holds its Third Plenum – a meeting that creates a roadmap for reforming the economy and society over the subsequent five years. The next such meeting should take place over the second half of November 2023. Experts speculate that China could announce a greater urbanization push (which could spur productivity and help the housing market), introduce measures to spur consumption and business sentiment, and seek a better balance between state-owned enterprises and the private sector. These could be moderately helpful for growth.
-With contributions from Vivien Lee and Aaron Ma