MacroMemo | 03 October 2023

Recession questions

We have been predicting a recession for quite some time. The possibility first came into focus in the spring of 2022, and we believed a recession could be imminent as long as a year ago. That raises two questions.

1. How has the recession proven elusive for so long?
   
   
2. Why is a recession still likely – in our view, anyhow – despite a year of surprising resilience?

One recession fended off

On the first matter – with regard to how a recession has been fended off for more than a year – the most useful way to think of it is that central bank rate hikes were never the main reason for a recession call in late 2022 and early 2023. They were part of the story, especially as negative wealth effects kicked in. But the lags involved were such that the peak effect of higher interest rates would theoretically arrive later.

Instead, the best argument for a recession a year ago was that there were so many other headwinds simultaneously blowing. High inflation wasn’t just a motivation for tighter monetary policy, it was also exerting a corrosive effect on economic growth all by itself. The war in Ukraine hurt trade flows and boosted commodity prices, creating an abrupt energy shock. Supply chain problems were still significant in 2022. China descended into recurring lockdowns across the year. North American home prices were in free fall.

This motley assortment of economic problems was theoretically sufficient to trigger a recession.

So why didn’t one happen? One reason is that the economy demonstrated impressive resilience – a recurring theme since the onset of the pandemic in 2020 and supported in significant part by government stimulus and pent-up demand.

But the big reason that a recession was avoided in late 2022 and the first half of 2023 was that many of those same headwinds started to fade, and fairly quickly:

• Falling inflation removed much of the corrosive effect on growth.
• The war in Ukraine sadly continues but the resulting energy shock is much diminished.
• Supply chain problems have since largely vanished.
• China’s lockdowns are over.
• North American home prices began to rise again in early 2023.

Recession still ahead?

So, with all of those problems significantly resolved and a recession handily avoided so far, why do we think a recession is still likely in the quarters ahead?

The main reason is that central banks have continued to raise short-term rates since last year. Longer-term bond yields have also continued to rise (see next chart). The U.S. 10-year yield is now at its highest level in more than 16 years (see subsequent chart).

U.S. yields rise on resilient economic data and Federal Reserve signal of higher rates for longer

^{As of 09/26/2023. Shaded area represents recession. Sources: U.S. Treasury, Macrobond, RBC GAM}

U.S. 10-year yields at highest level in more than 16 years

^{As of 09/26/2023. Shaded area represents recession. Sources: U.S. Treasury, Macrobond, RBC GAM}

The increase in the 10-year yield has substantially exceeded what one would normally expect in response to the central bank rate hikes. This is because the term premium has been actively rising and – at least according to the Federal Reserve Bank of New York’s estimate – is finally back in positive territory (see next chart). Many factors are helping to push yields higher, including:

• quantitative tightening
• the U.S. debt downgrade
• the large U.S. fiscal deficit
• rising Japanese and European yields
• China’s defense of its currency all push term yields higher

Term premium has risen recently

^{As of 09/26/2023. Sources: Federal Reserve Bank of New York, Macrobond, RBC GAM}

Lagged impact

Tighter monetary policy impacts the economy with a lag. In fact, we find the average lag from the first U.S. rate hike to the onset of a recession averages a remarkable 27 months (see next table). Thus, technically, the window for a monetary policy-induced recession is just opening. June 2024 would be the 27-month historical norm for the start of a recession given the U.S. Federal Reserve’s first rate hike in March 2022.

^{U.S. business cycles and tightening cycles. Sources: U.S. Federal Reserve, National Bureau of Economic Research, Macrobond, RBC GAM}

Reduced rate sensitivity overstated

While it is probably fair to concede that the U.S. economy in particular is somewhat less rate sensitive than in the past, this argument must not be advanced too aggressively. For one, the amount of monetary tightening and the pace at which it was delivered was sufficiently aggressive that it at least partially offsets the diminished sensitivity.

Second, it is critical to appreciate that higher interest rates impede economic growth through a variety of channels. Yes, the cost of servicing debt is one of them, and this effect is somewhat diminished given the aforementioned reduced rate sensitivity. Yet this channel is less all-encompassing than many imagine. There is always a party on the opposite side of every loan receiving additional income as rates rise.

Critically, the level of interest rates has a bearing on decisions about the future, regardless of the level of existing debt and the cost of servicing that outstanding debt. Companies contemplating whether to invest in a new factory must be confident that the return on their investment will exceed the elevated cost of funding it. For individuals, anyone wanting to buy a new vehicle or to enter the property market for the first time similarly faces high prevailing interest rates. To the extent economic output is the sum of businesses and individuals buying new things, higher interest rates continue to have a major effect in a fashion little different than prior cycles.

As such, we remain inclined to think that some are attributing economic resilience to reduced rate sensitivity when in reality they are simply underestimating the lags involved.

New headwinds

Even as last year’s headwinds faded, the economy now faces some new special-factor headwinds. Several, including recent auto strikes, restarted student loan payments, the expiry of a national childcare subsidy, and a possible government shutdown (now at least temporarily delayed), are admittedly temporary in nature and U.S.-centric. But they still warrant acknowledgement and, in combination, could represent a key negative turning point.

Auto strikes

The unionized workers of Ford, Stellantis and General Motors have been on strike for several weeks. Only select plants have walked out, but the number is mounting. The pressure to reach a deal is significant.

In the context of the high inflation of the past few years and a tight labour market, the union desires a large leap in wages. There has been some movement on both sides toward a resolution, but the wage gap is still tens of percentage points apart. If the strike is limited to something like six weeks, in line with the most likely resolution date articulated by the Good Judgement Open probability market, the economic damage should be fairly small – on the order of around 0.2% chopped off gross domestic product (GDP). But with the swirling number of strikers and an uncertain end date, this calculus is far from precise.

Furthermore, there could be secondary effects in the form of impacts up and down the supply chain. Higher car prices would be undesirable given the need for inflation to continue falling.

Stimulus expiry

We wrote about the resumption of U.S. student loan payments on October 1 in our last #MacroMemo. In a nutshell, about 44 million Americans must now start repaying an average of U.S.$503 per month on U.S.$1.8 trillion in accumulated student loan debt. This amounts to a 0.2% hit to personal income, with the potential for outsized financial damage to households with especially large student debt levels.

A U.S.$24-billion program created in 2021 to support struggling American daycare centres during the pandemic expired on September 30. The money went directly to the daycares rather than to parents, and so one cannot conclude that the 9.6 million children who benefited will suddenly find childcare unaffordable, nor that the average household will be financially worse off in the short run. But the Century Foundation think tank argues that as many as 3.2 million children could lose access to childcare as a result of the program’s expiry. That sounds high, but the point is that it may become incrementally more difficult for parents to secure childcare for their children. This in turn will likely increase the cost of childcare, reduce its supply, and possibly force some parents to drop out of the labour force to care for their children.

Government shutdown temporarily averted

Wrongfooting betting markets had assigned a 75% probability of a U.S. government shutdown on October 1. However, a stopgap funding measure was just passed at the last possible moment, bumping the problem down the road to November 18.

A potential wildcard is whether Republican House Speaker McCarthy can survive his decision to initiate the funding bills in defiance of a portion of his party’s wishes. If not, an even more adversarial approach may result. Financial markets are watching, with Moody’s threatening to join the other two major debt rating agencies by downgrading the U.S. sovereign debt rating from AAA to AA+ if a proper budget deal is not brokered.

Both parties had agreed to moderate spending constraints when the debt ceiling was lifted in late May, but an influential faction of the Republican Party now wants even larger cutbacks.

There is still a very real chance of a government shutdown on the new deadline of November 18 given the pressing need to pass a budget for the coming year and the large ideological divide that remains between Democrats and Republicans. The last shutdown, spanning late 2018 through early 2019, lasted 35 days and did real economic damage even though it was only a partial shutdown. One might expect a shutdown to subtract 0.2% from GDP for each week that it lasts, before the economy aggressively rebounds once the shutdown is resolved.

A tempering factor this time is that federal workers who are furloughed by a shutdown will be retroactively paid once the shutdown is resolved – in contrast to prior shutdowns when such wages were lost forever.

In summary, it isn’t so much that these problems by themselves are large enough to plunge the U.S. economy into recession. Instead, they could throw the economy sufficiently off balance to pull forward or otherwise trigger the long-awaited interest-rate driven recession.

U.S. 2024 election ahead

A quick first comment on the U.S. 2024 presidential election. It is still fairly distant, but coming into relevance, if not quite into focus.

There is still much that could happen, but at present it seems most likely to be a reprise of the 2020 election. That is to say, a re-match of Democratic Party President Biden versus Republican Party former President Trump. Opinions vary on who would win such a contest (see next chart). A weighted average of polls argues Trump is slightly ahead (RealClearPolitics), as does a betting website (oddschecker). Conversely, the probability market maintained by PredictIt argues that Biden is slightly more likely to win.

2024 U.S. presidential election appears close: Biden vs. Trump

^{As of 09/25/2023. RealClearPolitics poll averages Biden versus Trump matchups only. Other polls acknowledge possibility of other candidates contesting in the election. Predictit probability of winning derived from prediction markets data. oddschecker probability of winning derived from median daily betting odds. Sources: oddschecker, Predictit, RealClearPolitics (RCP), Macrobond, RBC GAM}

According to oddschecker, Biden led in a theoretical head-to-head matchup for the bulk of 2023, before abruptly relinquishing the lead to Trump in recent weeks (see next chart).

Trump has become the frontrunner in betting markets

^{As of 09/25/2023. Probability derived from median daily betting odds from oddschecker.com. Sources: oddschecker, RBC GAM}

It is too early to talk at any length about policy differences, though of course there are large potential divides with regard to the appetite for green initiatives (Biden), tax cuts (Trump) and Ukraine policy (Trump might provide less support to Ukraine). On the other hand, both appear to have few qualms about running large deficits, and both are anti-China.

It must also be mentioned that a Biden—Trump matchup isn’t quite assured. President Biden was already the oldest person to claim the Presidency at the beginning of his first term. Concerns about his age (80) have been raised with sufficient frequency that it is not impossible that he will be challenged by another Democratic Party candidate or opt not to run for health-related reasons. With regard to Trump, he is also quite old (77), and there are a large number of other candidates actively seeking the Republican nomination. Trump is leading by a considerable margin and has extended that advantage significantly over 2023 versus Florida Governor DeSantis. But it is not impossible that one of the also-rans catches fire, or that Trump’s various legal woes render him less viable or even ineligible.

Assorted economic developments

There continues to be a gap between U.S. and Eurozone economic data, with U.S. data proving relatively more resilient (see next chart).

Economic growth diverges in Eurozone and U.S.

^{As of 09/25/2023. Sources: Citigroup, Bloomberg, RBC GAM}

Market insight

Pessimists will note that the stock market has lately soured, with the S&P 500 drifting lower since the start of August after a happy appreciation that began in March. Compounding this, the earlier strength was never fully what it seemed, concentrated as it was in the “Magnificent Seven” tech stocks (see next chart).

U.S. equity indices are drifting lower

^{As of 09/26/2023. Magnificent-7 is a cap-weighted index which includes Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta and Tesla. Sources: Bloomberg, RBC GAM}

Conversely, optimists will note that credit spreads in the bond market, while not low, remain fairly tame and do not appear to anticipate near-term distress (see next chart). The credit market is far from perfect but is known as one of the more foresighted corners of the financial market.

U.S. credit spreads have narrowed

^{As of 09/26/2023. Spreads of ICE BofA U.S. Corporate Index. Shaded area represents recession. Sources: ICE, Bloomberg, RBC GAM}

U.S. Q3 GDP tracking

The third quarter of 2023 is now officially complete, though we will have to wait for some time before receiving an official estimate of economic growth for the quarter. As we noted in the prior #MacroMemo, there is the chance it is quite a robust figure, with the well-regarded Federal Reserve Bank of Atlanta’s GDPNow metric pointing to a remarkable 4.9% annualized gain. That would be more than twice the normal rate of economic growth.

However, this is not the only credible opinion on the subject. The New York Fed recently resumed publishing its own GDP Nowcast. It predicts a much more sedate 2.1% annualized gain. Finally, the St. Louis Fed’s GDP Nowcast points to just a 1.6% annualized gain.

This 3.3 percentage point gap is bizarrely large. It is hard to know what to do with it. The Atlanta Fed is clearly outnumbered, but it is also the most well-regarded of the three models. But, in fairness, this claim of superiority is no longer verifiable as the New York Fed reports that it has now enhanced its model.

As such, let us say that the uncertainty around the third quarter is unusually high. The outcome could be anything from decent to great. We economists will be watching with fascination for a hint as to whether the New York Fed’s new GDP Nowcast merits closer attention going forward.

Canadian Q3 GDP tracking

A similar exercise can be performed for Canadian Q3 GDP, albeit without the help of third-party GDP nowcasting models.

Canadian July GDP was just released at a flat 0.0% reading. Statistics Canada tells us that August GDP is tentatively tracking +0.1%. Our own calculation points to +0.07%, which is effectively the same thing.

Our in-house nowcasting model then articulates the barest of declines for September GDP. Adding these together, it sums to a tiny third-quarter GDP gain of +0.1% annualized. Coming on the heels of a slightly negative print in the second quarter, it is hardly a resounding endorsement of the Canadian economy.

Inflation backup

As expected, August inflation failed to provide a further deceleration of overall price pressures. Due to a mix of less friendly base effects and higher gas prices, the U.S. headline Consumer Price Index (CPI) rose from 3.2% year-over-year (YoY) to 3.7% YoY. This was slightly above the consensus, but in line with our own forecast.

The story was basically the same in Canada, though the leap was slightly greater, from 3.3% to 4.0% YoY. A difference between the two countries is the way shelter costs are captured in Canada, which includes mortgage interest payments. The cost of this has surged as mortgage rates have risen, generating a strongly positive inflation impulse that is preventing shelter costs from decelerating even as home prices themselves have eased (see next chart).

Higher mortgage rates lift Canadian shelter inflation

^{As of August 2023. Sources: Statistics Canada, Macrobond, RBC GAM}

Turning back to U.S. inflation, arguably the main problem with the latest inflation print was not that gas prices were higher – this was well understood in advance, much as it is understood that they will likely exert further upside pressure in the September data. Instead, the surprising development was that core inflation rose by 0.3% month-over-month (MoM), above the 0.2% MoM increases recorded in the prior two months. Even after we exclude gas prices, there was a touch of heat that didn’t previously exist.

Fortunately, a big part of the extra core heat came from airline ticket prices, which were merely reversing three months of notable weakness. Other key CPI components looked pretty good: shelter inflation rose by just 0.3%, the weakest in months (see next chart) and food inflation was a relatively subdued +0.2%. The breadth of inflationary pressures also continues to narrow (see subsequent chart).

In short, we would say that the underlying inflation story remains OK and is more consistent with gradual improvement than with a new problematic inflation spike forming.

U.S. dwelling costs in Consumer Price Index drop

^{As of August 2023. Shaded area represents recession. Sources: U.S. Bureau of Labor Statistics (BLS), Macrobond, RBC GAM}

High inflation in the U.S. is becoming much less broad

^{As of August 2023. Share of CPI components with year-over-year % change falling within the ranges specified. Sources: Haver Analytics, RBC GAM}

Hawkish pauses

Central banks have slowed their tightening greatly. Indeed, the U.S. Federal Reserve, Bank of Canada and Bank of England all left their policy rates unchanged at their most recent meeting, while the European Central Bank signalled that its September 25 basis point rate increase may have been its last.

All parties are heavily data dependent and given extraordinary uncertainty around the path for the economy from here, and whether inflation is truly vanquished, there are any number of plausible scenarios, ranging from further rounds of tightening to aggressive monetary easing.

But the most likely scenario is a period of time with policy rates gliding along the current relatively elevated plateau. Hawkish pauses are the order of the day. Central banks don’t want to raise short-term rates further, but they also don’t want to make any kind of concession toward near-term rate cuts lest the market get ahead of itself via lower bond yields, eroding some of the work central banks have done to cool inflation.

In the latest U.S. Fed decision, the hawkish aspect of the pause was on full display. Despite an unchanged fed funds rate and a barely altered statement, Federal Open Market Committee participants upgraded their GDP and unemployment forecasts, and indicated they are in less of a rush to cut rates over the coming few years. There were even a handful of participants who think the neutral policy rate may be higher than previously imagined.

We would counter that the neutral rate may not have increased as much as some think and that rate cuts usually happen more abruptly and aggressively than central banks initially intend (and certainly than they convey). But all of that remains speculative at this point, and it is premature to bet too heavily on rate cuts in the coming months.

^{With contributions from Vivien Lee and Aaron Ma}