Summary

There are a range of COVID-19 positives and negatives to share this week.

Negatives:

• The global daily infection rate continues to rise.
• Many U.S. states and a sizeable chunk of the emerging market world continue to record worsening virus numbers.
• The U.S. economy appears to be wobbling.
• Accordingly, we are in the process of downgrading our U.S. growth forecast.

Positives:

• The COVID-19 fatality figures remain surprisingly muted even as the infection numbers deteriorate badly.
• Arizona’s infection figures may be starting to peak – an important development if true, as Arizona has been a bellwether state.
• The economic recovery continues outside of the U.S.

Virus figures

The global COVID-19 trend remains challenging. New records are being set in the number of new daily infections, with a slight uptrend in the number of deaths (see next chart).

Spread of COVID-19 globally, cases and deaths

Note: As of 07/13/2020. 7-day moving average of cases & deaths indexed to 100. Source: ECDC, Macrobond, RBC GAM

All told, the world is now recording roughly 230,000 new infections per day. Nearly 14 million people have now been infected in total. More than 550,000 people have now died, with another 5,000-plus dying each day globally. It is some consolation that the equivalent fatality rate in April was higher, in the realm of 7,000 per day.

The contrast between the record number of infections and the tamer death rate can seemingly be explained by these factors:

• a lag in the fatality numbers
• additional testing that has captured a rising fraction of the infected
• a shift toward younger people being infected
• higher quality medical care.

This is not to say that all countries are in trouble. In fact, most developed-world countries are looking quite good right now, including Canada and the U.K. (see next two charts). Much of continental Europe is in a similarly enviable position. That said, few have continued falling over the past week. It is worrying that bars and indoor restaurants are now being opened in many jurisdictions, as that is believed to be a key vector of transmission in the U.S.

Spread of COVID-19 in Canada

Note: As of 07/13/2020. Source: ECDC, Macrobond, RBC GAM

Spread of COVID-19 in the U.K.

Note: As of 07/13/2020. Source: ECDC, Macrobond, RBC GAM

The U.S., conversely, remains at the problematic end of the spectrum. The country is recording a world-leading 60,000 cases per day and an aggressively rising trend (see next chart).

Spread of COVID-19 in the U.S.

Note: As of 07/13/2020. Source: ECDC, Macrobond, RBC GAM

As with the global figures, the U.S. fatality numbers have not deteriorated as badly as the raw infection numbers, though an upward trend is starting to become visible (see next chart). This could yet deteriorate further even if the infection rate were to stabilize given that the lag from an identified infection to death is around two weeks. Still, the fatality numbers remain much less distressing than they were in April – an important silver lining.

COVID-19 deaths in the U.S.

Note: As of 07/13/2020. Source: ECDC, Macrobond, RBC GAM

State developments

Most U.S. states continue to struggle, though the count with a transmission rate greater than one has declined nicely in recent days, from a high of 47 to “just” 30 today. That still means the majority of states are spiraling out of control, but no longer the virtual entirety.

Arizona is worth watching very closely. Although its raw infection rate isn’t especially notable, on a population-adjusted basis it has been among the most severely affected states during this second wave of COVID-19. Interestingly, the state’s trend infection rate is now starting to fall (see next chart), and the number of hospitalizations has seemingly stabilized. The state did make several important policy changes over the past several weeks, such that this could well represent a genuine peak. The next few weeks will be key in determining whether this pivot is an illusion nor not. If genuine, it could anticipate a similar peak in many other U.S. states that have also begun to get serious about the virus.

State of Arizona

Note: As of 07/13/2020. 7-day moving average of daily new cases used as trendline. Positive rate calculated as 3-day moving average of new cases/new tests. Source: The COVID Tracking Project, Macrobond, RBC GAM

Texas and Florida, on the other hand, continue to record ever-higher infection counts, with 10,000 per day in Texas and a spike to more than 15,000 per day recently in Florida (see next two charts).

State of Texas

Note: As of 07/13/2020. 7-day moving average of daily new cases used as trendline. Source: The COVID Tracking Project, Macrobond, RBC GAM

State of Florida

Note: As of 07/13/2020. 7-day moving average of daily new cases used as trendline. Positive rate calculated as 3-day moving average of new cases/new tests. Source: The COVID Tracking Project, Macrobond, RBC GAM

When will the U.S. tame the virus?

U.S. policymakers, businesses and households have begun to adjust to the country’s rising virus count. The question is how long it will take this adjustment to appear in the data, and whether the remedies will prove sufficient?

On the first question, Arizona may already be peaking, as previously detailed. But, in general, one would expect tighter rules to take two to three weeks to show up in the data if the March-April experience provides any guide. As such, if the actions taken so far prove sufficient, the daily infection rate could begin to decline again by late July.

With regard to whether the actions to date actually will prove sufficient, that is harder to say. The new restrictions don’t come close to matching the policies implemented in late March. So any pivot is unlikely to be so extreme that the virus count begins to fall precipitously.

To the extent that the new U.S. rules look increasingly like that of the rest of the developed world – widespread mask usage and more limited restaurant and bar access – the changes seems likely to stabilize the infection rate, but not necessarily to reduce it significantly given that much of the rest of the developed world is on an approximately sideways trajectory.

Reflecting the great uncertainty surrounding the issue, a recent internal poll found that the most popular single answer for when the U.S. virus count will peak was in the final quarter of 2020. Yet there were collectively more people who selected July, August or September, with August a particularly crowded option. Conversely, a few curmudgeons thought the virus wouldn’t peak in the U.S. until 2021. Time will tell.

Of course, even if the virus were to peak in a few weeks, this would not mean that the economy could be immediately reopened. With the first wave, governments waited several further weeks before beginning to incrementally reopen their economies.

Will other developed countries run into trouble?

We believe other developed countries will continue to succeed in limiting the spread of the virus.

This is not to say that they can continue to reopen their economies and social activities indefinitely. We are already nervous that they may have gone slightly too far.

But they appear to possess a greater capacity for nimble adjustment than the U.S. does, such that they should be able and willing to tighten their rules again should this prove necessary, without first delaying for a month. This speed and flexibility has already been demonstrated by Australia and Spain after recent outbreaks.

Emerging market challenges

Emerging market nations continue to suffer the worst of the COVID-19 spread, though not by much any longer (see next chart).

COVID-19 hitting emerging market countries now

Note: As of 07/13/2020. DM aggregates case count from France, Germany, Italy, Spain, U.K. and U.S., and represents 34.7% of global cases. EM aggregates case count from Brazil, India, Iran, Peru, Russia and Turkey, and represents 33.1% of global cases. Source: ECDC, Macrobond, RBC GAM

Brazil remains quite challenged. Its president even recently contracted the virus. But the reported infection rate has actually cooled somewhat recently (see next chart). It has now settled below the U.S. equivalent, at around 40,000 per day. Much of Latin America remains afflicted, with Mexico, Peru and Chile also notable.

Spread of COVID-19 in Brazil

Note: As of 07/13/2020. Source: ECDC, Macrobond, RBC GAM

India now records more than 25,000 new infections per day, though its enormous population means the virus is still much less pervasive than in many other countries (see next chart).

Spread of COVID-19 in India

Note: As of 07/13/2020. Source: ECDC, Macrobond, RBC GAM

Africa has not gone completely unscathed. The continent has now tallied a total of nearly 600,000 infections, with South Africa emerging as a particular problem spot, with 10,000 infections per day and rising (see next chart). The continent remains difficult to assess, as it has many great advantages in the fight against COVID-19: youth, less obesity, warm weather, and a population that spends a larger fraction of its time in the relative safety of outdoors. But Africa also has some unique challenges:

• malnutrition
• an economy that doesn’t lend itself to social distancing
• under-testing
• more limited medical capacity.

Spread of COVID-19 in South Africa

Note: As of 07/13/2020. Source: ECDC, Macrobond, RBC GAM

Economic developments

The U.S. economy is now stumbling, presumably in response to the rising infection count and tightening social distancing rules. Our economic activity index continues to point to a slight dip in activity after a long period of acceleration (see next chart).

Note: As of 07/01/2020. Economic Activity Index is the average of 10 weekly economic data series measuring the percentage change versus pre-COVID (year-over-year or versus a defined period before COVID-19 outbreak). Source: Bank of America, Goldman Sachs, OpenTable, Macrobond, RBC GAM

In fairness, the diagnosis isn’t quite as straight-forward as it looks. For one, not all indicators are actively declining, though most are. Even among those that are, it is hardly a linear affair – as demonstrated by the wobbles in a business sales survey (see next chart).

New orders and sales of U.S. businesses hammered by COVID-19

Note: As of 07/05/2020. Estimated as weighted average of % change in new orders or sales for all respondents. Source: Weekly Business Outlook Survey on the COVID-19 Outbreak, Federal Reserve Bank of Philadelphia, RBC GAM

And there is also the question of seasonality. Traditional economic data comes already seasonally adjusted. This is to say, the data has already been massaged to adjust for things like Easter landing on different months, a varying number of weekdays per month, summer jobs for students, and so on. But much of the real-time economic data we are now relying on does not make such adjustments.

As a result, a key question right now is the extent to which the recent apparent economic weakness in the U.S. might just be a consequence of the 4^th of July holiday, and more generally of people taking their summer vacations. Looking at the hours worked data (see next chart), you can see the effects of Easter and the May long weekend in the data. It could be that the latest drop is simply a version of that.

Impact on hourly workers

Note: As of 07/09/2020. Impact compares hours worked in a day vs. median for the same day of the week in January 2020. Source: Homebase, Macrobond, RBC GAM

However, a closer examination of the chart reveals that the decline in hours worked has seemingly been greatest in the states that are suffering the most severe COVID-19 outbreak. This suggests that at least some part of the decline is probably genuine as opposed to seasonal. Just what fraction?

The next chart calculates the extent of the decline in hours worked by state. One might hypothesize that the seasonal aspect is perhaps a 6% decline in hours worked (as per New York and Illinois, which have fairly low virus counts), meaning that the true decline in hours worked in Arizona, Texas and Florida is in the realm of -8%, -6% and -4%, respectively. At the national level, the decline is only slight, but nevertheless real.

Hours worked has declined more in states with more active COVID-19 outbreaks

Note: As of 07/07/2020. Compares percentage point change in hours worked from recent peaks in June 2020. Source: Homebase, Macrobond, RBC GAM

Good job news

U.S. jobless claims for the latest week reported a slight improvement from 1.4 million to 1.3 million initial claims. But the real story was in the continuing claims data, which saw a big 700,000-person decline in the number of jobless recipients from one week to the next. Unfortunately, this continuing claims data only runs through June 27, meaning it predates the subsequent downturn in the real-time data. This coming Thursday’s release may prove instructive in this regard.

U.S. job openings are also quite interesting. The number of job openings is unsurprisingly down sharply, from 7.0 million outstanding before the virus to 5.4 million now (see next chart, depicted as the ratio of job openings to employed people). But this is still a respectable level, as good as any reading recorded between the years 2000 and 2014. The pandemic has simultaneously created a job shortage and a labour shortage. The confluence of these two events can explained by the radical shift in the economy that has left millions of retail, tourism and food services workers unemployed while simultaneously creating a need for delivery workers and others in a handful of other spaces. Matching the newly unemployed with those new opportunities is far from straightforward given the challenges of pivoting sectors and the impediments created by government support payments, childcare challenges and the fear of getting sick.

U.S. job openings rate had been on a downtrend before coronavirus outbreak

Note: As of May 2020. Shaded area represents recession. Source: BLS, Macrobond, RBC GAM

Canadian data

Canada’s June employment report has now been released, and it pleasantly surprised with nearly a million new jobs after the creation of 300,000 in May. This means that 43% of the country’s job losses have since been unwound. Simultaneously, a special question in the survey revealed that another 800,000 Canadians who had previously reported working fewer than half their usual hours have now recovered a significant fraction of those hours.

The latest consumer spending data from RBC Economics has now been published. By this metric, consumer spending via credit and debit cards is now an astonishing 4% higher than it was a year ago (see next chart). This suggests consumers are doing just fine, though with the caveats that some of the increase likely represents a compositional shift away from cash, and a chunk of the spending is being artificially supported by government stimulus.

Consumer spending returns to growth in late-June

(year-over-year change in debit & credit card spending)

As of July 2020. Source: RBC Economics, RBC Data & Analytics

Look ahead

Looking to the week ahead, two things seem particularly notable:

1. U.S. retail and industrial production will be released for June, providing further insight into the extent of the U.S. economic recovery before the country ran into a wall of COVID-19.
2. Chinese second-quarter GDP will be released, with the consensus looking for a remarkable +3% year-over-year performance. It is amazing that the Chinese economy is expected to be larger than it was a year ago. Granted, this pales compared to China’s desired growth rate of +6%. But it provides hope that the recovery might be a nimble one, at least for countries that have followed China’s lead in significantly reducing the presence of COVID-19.

Second-round damage

We continue to muse now and again about the second-round economic damage brewing beneath the surface. This is entirely separate from the second wave of COVID-19 presently afflicting the U.S. The second-round damage refers to the fact that, although the economy has staged a welcome recovery, it remains smaller than normal. This means that workers are still earning less than normal and company profits are still lower than usual.  Losses are still accruing each month. Problems may yet arise from this, most obviously of an insolvency nature.

A recent academic study identifies another variety of second-round damage. The authors estimate that 42% of the jobs lost due to the pandemic will not come back. That said, they simultaneously estimate that for every 100 job losses, there should be around 30 new hires. While the two figures don’t quite fit together as neatly as they look like they should, one might conclude without being too radically off base that the labour market may ultimately struggle to find a home for the residual between these two figures – or around 10% of those who have lost their jobs.

Will schools re-open?

A key question in assessing whether the economic expansion can continue into the fall is whether schools will be able to fully accommodate their usual student body. This is a challenging issue, for a number of reasons, including:

• The lives of children are viewed by society as the most precious of all.
• There is undeniably some risk involved in returning children to school.
• School boards are not well equipped to judge what constitutes an acceptable risk.

As a result, there has been a great deal of discussion about resuming virtual schooling in the fall or adopting a hybrid approach that combines virtual and physical schooling.

If pursued, these options would seriously interfere with the economic rebound: it will be very hard for parents to return to work if their children are not physically in school. While this issue affects all parents of school-aged children, it also threatens to increase inequality across gender and income lines.

To the extent children are not able to attend physical schools and their parents must work, daycare will no doubt play a role. But daycares are arguably even more pressed for space, and children will be left at least as exposed the virus, minus the education. Grandparents are a childcare option for some, but older people are the most vulnerable to the virus.

From a long-run perspective, virtual schooling is unlikely to convey the same amount of human capital onto the children, leaving them less prepared for the working world. Many children have already fallen significantly behind. From a social capital perspective, the virtual options fall even further short. 

The evidence available so far argues that children are much less affected than older cohorts by the virus, and may also be less capable of transmitting it to others. As such, there is a strong case to be made in returning to full-time schooling, at least in locations not presently in the jaws of the virus. A handful of countries have already done this, mostly to good effect. Camps this summer will provide another testing ground. Daycares have been open for the children of emergency workers throughout the pandemic. While the experience has not been perfect in all cases, neither have these proven to be hotbeds of infection.

Creative (and potentially crackpot) solutions for schools this fall include:

• Have Plexiglass shields between students.
• Have high school-aged children learn virtually, and use the empty high schools to spread out the elementary school children into additional classrooms.
• Use libraries, community centres, gyms, parks and school yards for instruction to reduce student density.
• Leave school windows at least partially open in the winter – even a limited dose of fresh air appears to significantly reduce transmission.
• Recruit retired teachers to teach some of the online classes.
• Have the board’s top high-school teachers record lectures used by all students, with the remaining teachers used to provide more personalized instruction and assistance as needed. Some subset of the high-school teachers could then be repurposed to the elementary system.

Whatever governments decide on the education file, the decision will be consequential for the economy both in the short run (allowing parents to work) and the long run (educating the next generation of workers).

Economic forecast update

Two months ago, we revised our forecasts moderately higher for two reasons: the trough of economic activity turned out to be slightly milder than we had assumed, and the economic recovery came together more quickly than we had imagined. Since then, there hasn’t been a pressing need to adjust our forecasts.

However, several things have recently transpired that oblige a further refresh.

1. We are increasingly able to map our mobility data and real-time economic data onto GDP. The uncertainty was initially too high to do this properly. But now that traditional economic data is available through May, it is possible to calibrate the decline and subsequent partial rebound in the non-traditional data to determine what it says about the economy right through to early July.

2. With specific regard to the U.S., we now see evidence that the U.S. recovery is fizzling. That needs to be factored into our forecasts, as previously we had penciled in a fairly steady recovery.

3. In a change that we have hinted at for more than a month, we are increasingly of the view that a full return to economic normality will take longer than previously imagined. This is simply because some sectors aren’t going to come back any time soon, and recessions are usually followed by multi-year rather than instantaneous recoveries.

Putting these three sets of thoughts together, we have tentatively downgraded our U.S. 2020 GDP growth forecast from -7.1% to -8.0%. To be clear, this may yet be tweaked. But it provides a sense for the direction and approximate scale of the appropriate adjustment.

The new forecast concedes that the U.S. economy is unlikely to recover half of its lost ground by the middle of July. It nearly pulled this off, with a big 45% of the loss having been reclaimed by late June, but output is again in retreat and we assume it shrinks back to 35% by the end of July before starting to inch higher again.

As a result, it now takes until December for the U.S. economy to reclaim half of its lost output, and until April 2022 to reclaim its prior peak (this was previously December 2021). The full return to a normal level of activity doesn’t happen until the middle of 2023 (previously the middle of 2022).

We have not yet completed the same exercise for other countries. Most do not enjoy so rich a set of real-time economic indicators as the U.S., though we can cobble together some approximation of the trajectory from each country’s mobility data. It isn’t yet clear whether forecast upgrades will prove appropriate, nor whether any countries will actually wind up with outright better 2020 growth forecasts than the U.S., but this looks to be a distinct possibility. At a minimum, some convergence between the U.S. and the other countries is likely.

Optimizing the pandemic response

The key to controlling COVID-19 is to limit the number of times that humans interact with one another. There are many ways to do this. If it is deemed crucial that people be allowed to go to bars and indoor restaurants, this could well prove sustainable so long as visits to offices and places of worship are limited, for instance.

A recent Brookings Institute report sheds some interesting light on this subject. It evaluates two options: shutting down the economy versus shutting down non-economic interactions. While either would work to some extent, their modelling argues that the virus control effort would enjoy fewer than half as many cumulative deaths by allowing the economy to run free while greatly limiting non-economic interactions, rather than the reverse. Thus, there are ways to get the great bulk of the economy up and running and also control the virus, but only if one is willing to sacrifice such social activities as church, sports, theatre, live entertainment and even extended family gatherings.

For the moment, it doesn’t seem that society is willing to sacrifice such things. But the option exists should a serious second shut-down prove necessary. While we are used to debating the relative importance of human lives versus the economy, it would appear there is now a third variable to consider: the relative importance of social activities.

Bending the curve versus eradicating the virus

We now have sufficient information to attempt to assess whether countries should merely seek to “bend the curve” – flattening out the rate of transmission – or instead seek to eradicate the virus altogether. The latter obviously sounds superior, but it must be weighed against the economic damage associated with the effort.

The margin between the different approaches can be quite small. A transmission rate of 1.1 would increase the number of daily infections 141-fold over the span of a single year. Conversely, a transmission rate of 0.9 (eradication) would reduce the daily infection rate by more than 200 times within the span of a year. Lying in the middle, a transmission rate of 1.0 (flattening the curve) would of course leave the number of daily infections exactly unchanged.

As we examine the U.S. mobility data trajectory, it seems to us that the country was on a sustainable path until early June, at which point the transmission rate began to drive above the critical threshold of one. This suggests that the limit for U.S. mobility was breached sometime around then.

In fairness, identifying the delineation point between an eradication strategy and bending the curve is a moving target, as people, businesses and governments are getting ever better at learning how to function safely via a two-metre distance, mask wearing, better testing, Plexiglass shields, and so on. More can be done now than was possible a few months ago.

Nevertheless, our rough-hewn attempt yields the conclusion that an eradication strategy would require a level of GDP about 3% smaller than a curve-flattening strategy. In the case of Canada, that represents the loss of USD$51 billion of economic output over a year.

At the current Canadian fatality rate of around 20 deaths per day, a curve-bending strategy would yield around 7,300 deaths per year, whereas an eradication strategy would result in just 1,394 deaths over the first year. The question is then whether the extra 5,886 deaths are worth USD$51 billion. It turns out that one would have to value each life at greater than USD$8.7 million for this to be a worthwhile exchange. That sounds astonishingly high, but it is actually quite close to the standard value placed on a life by the U.S. government. As such, while it sounds horribly cavalier to say, policymakers should be roughly indifferent between pursuing an eradication or a curve-bending strategy with those parameters.

However, there is an extra consideration. That math only works if you imagine the virus then disappears in a year’s time. This is not a bad assumption as vaccines are expected to arrive around then. However, if you think that the virus might stick around for several years, this tilts the argument toward pursuing the eradication strategy, at least for the first year. After the first year, the number of deaths per day should be down to nearly nothing, allowing a shift toward the curve flattening strategy thereafter (and reclaiming the lost 3 percentage points of economic output in subsequent years). To the extent a further 5,886 lives would be saved each subsequent year, the financial math then becomes quite favourable toward the eradication strategy. For our part, we believe a solution is likely to come over the next year, leaving the optimal strategy ambiguous.

What about Sweden?

A related debate has raged over the approach that Sweden has taken: allowing the virus to spread more freely via fewer economic restrictions.

A handful of newspaper articles have concluded that Sweden derived no economic benefit at all from its strategy, but this is not quite true. Most forecasts have Swedish GDP on track to decline by between 5% and 7% in 2020, whereas most of its neighbours are on track for declines of between 5% and 10%. The country appears to have squeezed out perhaps a 2 percentage point GDP advantage. Similarly, consumer credit card data shows a 29% drop for Denmark versus a 25% drop for Sweden – another slight advantage.

Of course, it is another question altogether whether this slight economic boost is worthwhile given Sweden’s additional deaths. The country has suffered a death rate of 54 per 100,000 people, versus 5 for Norway, 6 for Finland, 11 for Denmark and Germany, and (for further context) 24 for Canada. To the extent this represents less than half a year of COVID-19 deaths (though they don’t come at a constant rate), we might guess that Sweden will suffer an extra 10,000 deaths from its strategy relative to its neighbours. This computes to an economic benefit of USD$1.1 million per extra death. Given the earlier discussion, this was probably not a worthwhile trade-off for Sweden. It becomes even less so if the pandemic lasts for longer than a year. Sweden probably should have locked down more.

The only caveat in this analysis is that if a vaccine proves elusive, herd immunity may be the necessary strategy. In that case, Sweden could be said to have secured a head start in that pursuit because the deaths would be inevitable in that scenario.

Deficits and taxes

The fiscal environment will be challenging in the coming years. Countries are running huge deficits, with the U.S. 2020 deficit on track to be in the range of $3 trillion. Canada’s 2020 deficit is now projected to be an unprecedented $343 billion (16% of GDP!).

The International Monetary Fund (IMF) and others have already projected 20 percentage point-plus increases in public debt-to-GDP ratios for many developed countries.

Canada’s bulging debt load was viewed as being significant enough that ratings agency Fitch downgraded the country from a treasured AAA to AA+. To be fair, the country still effectively maintains its AAA rating as the two other major ratings agencies have not altered their judgement. Furthermore, most countries have long since lost their AAA rating and bond markets appear not to care a whit, in sharp contrast to the last time this happened this happened to Canada in the 1990s. As such, little will likely befall Canada as a result of the move, but the shift is at a minimum symbolic of a more challenging path ahead.

The key question in all of this is whether taxes will have to go up to pay for the larger deficits. There have been a variety of politicians and political pundits who have opined on the subject, with a significant fraction suggesting that higher taxes may prove necessary. Far be it for us to reject this idea, particularly when politicians are the ones who will ultimately determine what happens, but it seems well short of a 100% likelihood.

For higher taxes to prove necessary, one of four things would have to happen:

1. Special government stimulus programs would have to be made permanent. This isn’t impossible, but we don’t expect the major programs to become permanent expenses, in part because most are explicitly temporary and in part because even if they weren’t, demand would naturally shrink as the economy recovers. While some of the new programs have opened our eyes to what can be done to minimize suffering during a recession and/or to address chronic poverty, the programs have also been shown to have flaws – some recipients have reportedly been slow to return to work because the benefits are more attractive than working.
2. There would have to be a serious will to pay down the additional debt that has recently accumulated. We are profoundly dubious of this. That simply hasn’t been the way that governments have operated in recent years. Voters don’t reward this behavior. It is highly unlikely that governments later decide to run giant surpluses so that they can pay off today’s enormous deficits.
3. Debt-servicing costs would have to become unmanageable. But this seems unlikely given that we are now operating in what appears to be a structurally low interest rate environment. Our calculations suggest the debt-servicing burden will inevitably rise slightly given the additional debt set to be on government books. But the burden will not rise to the point of needing a significant new revenue source.
4. The economy would need to remain mired in recession or at least to seriously underperform for many years. While we are increasingly attuned to the idea that it will take several years for economies to return to normal (we now believe the U.S. economy won’t fully normalize until the middle of 2023), governments shouldn’t need a permanent new revenue source unless there is set to be a permanent large deficit. This shouldn’t happen unless the economy materially undershoots expectations. And note that even though it took the U.S. economy nine years to fully normalize after the global financial crisis, taxes fell rather than rose over the period.

Furthermore, it would be quite unusual to be hiking taxes while emerging from a recession. There will already be a fiscal headwind at that time as government stimulus programs fade, and politicians will still be anxious about restoring the economy to its full capacity. Reflecting this, the U.K. is now musing about tax cuts rather than tax increases.

As such, we don’t believe higher taxes are automatic over the next few years.

To the extent there is a risk of higher taxes, it has less to do with COVID-19, and more to do with the political environment. Canada’s minority government skews to the left, making tax hikes conceivable. In the U.S., betting markets now argue for a Democratic Party sweep in November – also an environment that could well see higher taxes for reasons unrelated to the pandemic.

What might higher taxes look like, if they did arrive? Despite considerable discussion, a higher sales tax seems unlikely to the extent it is a famously unpopular tax with voters and inadvertently targets low income households and retailers. Conversely, despite its economic inefficiency, a higher tax on investment income seems conceivable, as would a new, higher bracket targeted at the extremely wealthy. In the U.S., a Democratic administration might well aim to increase the corporate income tax rate as well.

But, again, it is far from clear that COVID-19 necessitates higher taxes by itself.

-With contributions from Vivien Lee and Kiki Oyerinde