Global Investment Outlook

Our latest quarterly Global Investment Outlook is now available, including an economics article entitled Making progress amid new risks.

Summary

The COVID-19 news is arguably tilting in a negative direction.

While the fatality figures remain fairly low and the economy is continuing to advance, the raw virus numbers are again deteriorating. The U.S. may now be deteriorating. The rate of economic growth could be about to wobble as efforts to combat the virus intensify. Vaccine prospects have slipped slightly (though still remain fairly good). And the world’s poorest people have been disproportionately hurt by the pandemic.

Virus developments

The number of new global COVID-19 infections per day continues to set new records, reaching 300,000 (see next chart).

Spread of COVID-19 globally

Note: As of 09/18/2020. Spike on 02/13/2020 due to methodology change. Source: ECDC, Macrobond, RBC GAM

The number of cumulative cases is approaching 30 million, while the number of cumulative deaths is on the cusp of one million. Fortunately, the daily fatality numbers are not themselves setting new records. They are even in slight decline, for reasons discussed later.

The majority of new cases remain in emerging market countries. But India – the worst affected among them in terms of the sheer numbers of cases – is tentatively flattening after a long increase (see next chart). Let the record show that the country fooled us earlier when its infection figures briefly stabilized in mid-August. But this flattening is already more enduring than the prior one. As such, there is a real chance that the Indian figures will begin to improve in the coming weeks.

COVID-19 cases and deaths in India

Note: As of 09/18/2020. 7-day moving average of daily new cases and new deaths. Source: ECDC, Macrobond, RBC GAM

Latin American infection figures remain quite high, but are mostly improving in such hot spots as Brazil, Peru and Mexico. Elsewhere, China remains a particular bright spot. South Africa’s numbers are greatly improved. But Iran and Russia are again deteriorating and some Eastern European countries such as Poland are now recording rising numbers of infections.

Europe remains ablaze with its second viral wave. France now records 11,000 new cases per day and the numbers continue to rise unabated. Spain has a similar number of new daily cases, though the rate of increase is finally slowing and the country has announced a new set of lockdown measures (see next chart).

COVID-19 cases and deaths in Spain

Note: As of 09/18/2020. 7-day moving average of daily new cases and new deaths. Source: ECDC, Macrobond, RBC GAM

The U.K. continues on its undesirable upward trajectory, now reaching approximately 4,000 new cases per day and still rapidly rising.

Canada is similarly suffering mounting problems. The country is reporting around 1,000 new cases per day after having achieved a trough of around 300 to 400 cases per day over the summer (see next chart). Notable increases are visible in Ontario, Quebec and British Columbia.

COVID-19 cases and deaths in Canada

Note: As of 09/18/2020. 7-day moving average of daily new cases and new deaths. Source: ECDC, Macrobond, RBC GAM

The U.S. had until recently been on an ameliorating trajectory, having tamed its second wave. However that trend has now halted, and is – if anything – getting a bit worse again (see next chart). This is unfortunate as the U.S. was the shining light among developed nations in terms of successfully taming its second wave without too much economic damage.

Furthermore, even if the U.S. numbers fail to rise again from here -- despite the challenges of re-opened schools, cooler weather and re-opened business sectors -- the absolute number of cases remains far from ideal with around 40,000 new infections per day.

COVID-19 cases and deaths in the U.S.

Note: As of 09/18/2020. 7-day moving average of daily new cases and new deaths. Source: ECDC, Macrobond, RBC GAM

Second wave

As second waves occur across most of the developed world, the process of easing social distancing restrictions is beginning to reverse. Continental Europe is already well underway in this reversal.

Now, the U.K. and Canada are beginning to tighten their rules.

The U.K. is reducing the size of permitted social gatherings to six people, introducing a generalized 10 p.m. curfew in Northeast England, and advising people to get out less and to interact with one another less. The region’s schools remain open, as do all major business sectors. A prominent panel of British scientific advisors have proposed a second national lockdown to coincide with fall break for students. Yet the prime minister has so far rejected any proposal to shut down substantially.

In Canada, the provinces of Ontario and Quebec have done something very similar to the U.K.: limiting social gatherings to smaller sizes than before in certain regions. Quebec has also ordered the closure of bars, restaurants and casinos by midnight.

In the past, we have described policymakers as trying to juggle three competing objectives:

1. minimizing the number of infections
2. maximizing economic output
3. maximizing the amount of social interaction possible (this can be thought of as “fun”).

Naturally, there is considerable trade-off among the three objectives. As problems mount with the first, sacrifices are unavoidable in the latter two domains. For the moment, these governments are mainly opting to limit social interactions, allowing economic output to continue undisturbed.

An exception is British Columbia in Canada, which recently closed its night clubs and banquet halls.

At this point, it is not clear that governments have done enough to tame the second wave. It could be that the mounting COVID-19 news persuades the public to be more careful – a voluntary solution. It could be that governments later opt to double-down on limiting social interactions. Or – as we suspect could prove most likely – they may have to turn their gaze toward explicitly limiting the riskiest economic activity: a financially costly move, but unavoidable if all else fails.

Testing struggles

It has become harder to secure a COVID-19 test in many parts of the developed world. In some cases, the amount of testing began to ebb over the summer, reflecting a reduction in supply.

But the issue is mainly a function of increased demand for testing:

• There are a rising number of people infected by COVID-19, and by extension even more people who have been in contact with those who are infected.
• Increasing socializing over the past several months means that the average sick person was in contact with more people today than in the early stages of the pandemic, requiring more follow-up testing.
• The return to school creates another possible vector for transmission, and when one student tests positive, the others in their cohort and perhaps even across their school must be tested.
• Many governments have encouraged anyone to get tested, even when people have no symptoms and no contact with any infected person. Governments may need to scale back on this messaging.
• Cold and flu season is beginning, meaning that many people will soon have COVID-like symptoms.

Fortunately, the amount of testing is now rising in many jurisdictions, but significant further increases will likely be necessary. Accurate real-time results would represent a huge leap forward. However, as yet the technology is imperfect.

Cases versus deaths

We noted earlier that while the number of daily infections are setting records, the number of daily deaths are not. Canada, for example, has half as many infections as in May, but 50 times fewer fatalities. Other countries report similarly enormous divergences.

This presents two obvious questions: why has this gap opened up, and should it change the policy response?

Why has this gap opened up?

The fatality figures have not increased in lock-step with the case numbers for several reasons:

• The fatality numbers are known to lag the infection numbers by several weeks, though this is not the central reason for the wedge.
• The first infection wave was almost certainly undercounted to a sizeable degree. This is to say, the first wave was likely much greater for most countries than the second wave, even when the official infection numbers suggest otherwise. As such, it makes sense that the fatality numbers – which were less distorted throughout – are comparatively better this time.
• The second wave is tending to infect younger people. Older and more vulnerable people are still being well protected. This makes a massive difference. COVID-19 is hundreds of times more fatal for those over the age of 60 relative to those under 40, let alone the divide between someone who is truly elderly and truly young.
• The quality of medical care has increased significantly. Not only are hospitals less overloaded than they were in the spring, but treatment has been considerably refined, with better procedures and also better drugs and medicines, including the use of remdesivir and dexamethasone – which each reduce the probability of death by 30% to 50%.

We speculated a few weeks ago that if the original true fatality rate was 0.5% to 1.0% – as several credible studies have argued – then the improvement in medical care alone should theoretically have reduced the fatality rate to something like 0.1% to 0.5% – two to five times lower than before for the same cohort.

Should this change the policy response?

It is fantastic news that the fatality rate is likely lower than before, and therefore that the reduced number of fatalities in the second wave is likely genuine.

However, this still leaves COVID-19 markedly more deadly than the flu. Furthermore, it is more easily transmitted than the flu, and also appears to do serious and potentially lasting damage to many of those who survive its initial assault. As such, it remains something worth working very hard to avoid.

So, should this change how policymakers respond to the virus? It undeniably presents an additional argument in favour of opening the economy and pursuing herd immunity. But the best strategy likely remains containing the virus, especially given that such efforts do not appear to require a complete economic shutdown, and given that a vaccine will likely become available over the next few quarters.

Does the U.S. have fewer deaths than usual?

We have lately encountered a number of reports citing a chart that purports to show a collapse in the number of deaths in the U.S., to the point that the U.S. appears to be recording substantially fewer deaths than normal right now (see next chart).

Unweighted – not accounted for underreporting

Note: As of the week ending 09/05/2020. Data in recent weeks are incomplete due to delay of submission of death records to NCHS for up to 8 weeks or more. Source: CDC, Macrobond, RBC GAM

This is indeed real data from the U.S. Center for Disease Control. However, we are quite skeptical that the U.S. is now managing outright fewer deaths than normal even as a pandemic unfolds.

The key to understanding this is that the chart is constructed using back-dated data, aligned to when each person originally died, not when the death was tabulated. Given that the paperwork associated with fatalities results in a reporting delay of up to eight weeks, this means that the most recent eight weeks of data are – at all times – underestimating the true numbers. In other words, you’d expect the number of reported deaths – and thus excess deaths – to appear to be collapsing at the very end of the series, at all times, only to be corrected later as lagged data arrives.

The one blurry contradiction in all of this is that the CDC indicates it makes an adjustment so as to minimize this bias. So might the sharp decline be genuine after all? It still seems highly doubtful, for several reasons:

• The data series aligned perfectly with expectations up until around two months ago – the point at which the theoretical distortion comes into play.
• We know that there are more than 5,000 COVID-19 deaths each week, so all else equal, one would expect there to be moderately more deaths than usual.
• While one can mount a credible argument that non-COVID-19 deaths could theoretically be lower than usual – due to fewer bleary-eyed commuters having car accidents, fewer deaths from the flu, etc. – if this were somehow saving 20,000 lives per week as per the final data point in the chart, shouldn’t even more lives have been saved earlier in the year when the country was truly locked down and car-driving was at its minimum? Instead, we find that the number of non-COVID-19 deaths was quite stable and normal looking throughout – up until the last two months.
• We have now watched the revisions to this data series for a week. Last week, the number of reported deaths for the week of August 29 was 46,079. This week, the same estimate has increased to 54,725. It will likely continue to increase over the next several weeks, eventually alighting upon a rational-looking 60,000 or so.

While we are knee-deep in this CDC data, it is worth sharing that the U.S. has experienced 225,000 excess deaths since COVID-19 arrived. This is actually slightly higher than the 199,000 COVID-19 deaths officially tallied. We suspect much of the difference represents COVID-19 deaths that were not captured as such due to the many ways that the virus can strike.

How is COVID-19 transmitting?

This has always been a big question for us, and never one that has been answered completely satisfactorily. Are the people who are getting sick mostly front-line health professionals? Are they front-line service workers? Are they people who ignore mask and social distancing recommendations? Or are they just unlucky people who happen to encounter the wrong person at the wrong time despite using appropriate caution?

We still don’t have great answers, in large part because many people presumably have no idea how they were infected. Many would have been infected by people who were either asymptomatic, minimally symptomatic or pre-symptomatic. To the extent one’s own symptoms don’t manifest for a week or longer after infection, it can require reconstructing one’s activities from many days before – a challenging task.

What we are able to say is based mostly on laboratory research and studies that evaluate statistics rather than tracing the contacts of infected people:

• Transmission via objects appears to be minimal.
• Face-mask usage is associated with diminished transmission.
• The duration of contact with an infected person is a key determinant – a brief period of time is much less dangerous than a lengthy encounter.
• A recent Lancet journal article found a strong correlation of 0.7 between the mobility data for each U.S. county and the county’s COVID-19 case growth. In other words, places where people were moving around a lot were more likely to suffer an outbreak. This seems logical.
• As we cited last week, a recent CDC analysis found that while there was an increased risk of contracting the virus among those who reported having attended a gathering of more than 10 people, took public transit or had a household member with COVID-19, the strongest predictor by far was whether the person had patronized an eat-in restaurant or bar. It is frankly astonishing that this was a better predictor than if someone within COVID-19 was living within the person’s own household.
• Illustrating that little can be said with absolute conviction, and seemingly contradicting the prior study, the province of Quebec recently justified its latest tightening efforts via a comment from the health minister. He stated that private gatherings and events like barbecues and weddings were the cause of much of the community spread in the province, whereas bars and restaurants had not been responsible for a significant number of infections.
• Young people are now transmitting the virus more than older people. It is unclear whether this is because they are being less responsible (perhaps recognizing their lesser vulnerability), or instead because they are the largest cohort of people (the single largest 5-year age bracket in the U.S. is aged 25 to 29, contrary to common belief), because people of that age are naturally the most active and social, or because they are more likely to work in front-line service jobs that may transmit the virus more easily.

Stalled progress

The Gates Foundation recently released its latest annual Goalkeepers Report, reviewing the state of progress toward the United Nations’ Sustainable Development Goals. Most years, cheery progress is reported in the number of people suffering extreme poverty and in various health goals. Indeed, even as the global economy grew only sluggishly over the past decade, the improvement in the quality of human life as measured in these gauges was quite heartening.

Alas, after years of progress, many of these measures have now suffered a serious deterioration at the hands of COVID-19. Bad news includes:

• Extreme poverty is up by 7% due to COVID-19. This means another 37 million people have fallen below the extreme poverty line of US$1.90/day of income. Income inequality has also increased.
• The vaccine coverage of children (not for COVID-19 but for a myriad of other conquered diseases) has retreated by the equivalent of 25 years to 1990s levels. This is considered a proxy for the overall functioning of the health care system.
• The incidence of malaria is predicted to increase by 10% after flat readings for several years.
• The poorest countries have much less room to help their economies: stimulus funding averages 22% of GDP in G20 countries, versus just 3% in sub-Saharan African countries.
• Globally, the fraction of children graduating from primary school who will have achieved a minimum proficiency in reading is expected to fall by 5 to 10 percentage points.
• If the wealthiest countries buy the first two billion doses of COVID-19 vaccine rather than distribute them more equally, nearly twice as many people may die from COVID-19.

This is depressing stuff. There is reason to think that much of the slippage can be undone over the next several years, but it is a serious setback nevertheless.

Falling mobility

Recently, the Apple mobility indicators suffered a fairly sharp decline (see next three charts).

Apple Mobility Trends, Spain

Note: As of 09/15/2020. Source: Apple Mobility Trends Reports, RBC GAM

Apple Mobility Trends, U.S.

Note: As of 09/15/2020. Source: Apple Mobility Trends Reports, RBC GAM

Apple Mobility Trends, Canada

Note: As of 09/15/2020. Source: Apple Mobility Trends Reports, RBC GAM

Why are the mobility measures in decline?

1. The decline in mobility may simply reflect a rising virus count and tighter government rules. This seems especially plausible for Spain. However, the U.S. was enjoying improving COVID-19 numbers up until very recently, and has not tightened its restrictions. Moreover, the decline in U.S. mobility is sharper than it was during the second wave itself. And although Canada’s virus numbers have been deteriorating subtly for some time, they have not elicited panic and the policy response has been fairly ginger and quite recent. As such, while the virus likely explains part of the retreat, we believe other factors are also relevant.

2. The summer holidays have now ended in the northern hemisphere and people have shifted back to work and school. It wasn’t obvious in advance whether this would increase or decrease mobility – do people drive more while on holiday or while going to work? – but it seems plausible to conclude that people drive more while on holiday, and the summer is known as the “driving season.”

3. The weather is now cooling. It makes complete sense that people would be going to the beach and the park less, and perhaps walking less in general. Of course, this doesn’t necessarily mean that the economy was weaker than before.

In short, we are not convinced that the economy is suffering overly, even as the mobility numbers go down.

Traditional economic data is usually seasonally adjusted: manipulated so that summer students quitting their summer jobs don’t pummel the employment rate, and the holiday shopping season doesn’t send retail sales to the moon and then back.

But we can’t seasonally adjust this new mobility data, or for that matter much of the real-time economic data. To do so properly would require many years of historical data that just don’t exist. To do so in a jerry-rigged fashion would require at least a year of normal data against which to compare the COVID-19 era, and that also doesn’t exist, for the most part. Perhaps the seasonal factors from traditional indicators can be redeployed for real-time indicators, though the effort would be highly imprecise.

The seasonal distortions didn’t initially matter as the scale of the economic contraction and then the economic rebound was so great that it dwarfed any seasonal aspect. But that is no longer the case now that the recovery has settled down to a more pedestrian clip.

Going forward, it appears we will have to rely to a greater extent on traditional economic indicators. This is not a great sacrifice in that economic data is no longer gyrating as aggressively as before. But it is unfortunate, especially as second and perhaps even third virus waves emerge that would ideally receive closer scrutiny.

Fortunately, there are a few ways to get a high-frequency take on the U.S. economy, at a minimum. For instance, U.S. initial jobless claims fell in the latest week to just 860K (from 893K), the lowest reading during the pandemic. Continuing claims were even more impressive, falling from 13.5 million to 12.6 million – an unemployment decline of nearly a million people.

Combined with a rising consensus growth forecast – visible both in the latest U.S. Federal Reserve decision and in the latest Consensus Economics outlook in which 10 of 13 major developed countries enjoyed an upgrade to their 2020 outlook – it would appear that the recovery is proceeding even as the mobility data swoons.

Fiscal update

U.S. fiscal

The U.S. has already gone over several fiscal cliffs, first when lump-sum cheques were issued in April (meaning that household incomes then fell sharply in May), and most recently when the country’s enhanced unemployment insurance program was scaled back at the start of August.

All told, U.S. fiscal support has shrunk by around $200 billion per month from its peak in April, though the level of household income – this is to say, organic household income plus government support – remains in line with pre-pandemic levels, albeit around 13% less than its brief apex.

While fading fiscal support represents a potentially profound economic drag, we have not seen too much damage, so far. For instance, U.S. retail sales managed to eke out a further increase in August even as the average unemployed person lost several hundred dollars per week of income.

Some offsetting economic help may have come from unemployed people opting to return to the workforce as their UI benefits became less attractive. But a survey by the Bipartisan Policy Centre found that only 16% of people who were receiving benefits but not looking for work indicated that this was because their benefits were greater than their prior paycheque.

Looking forward, negotiations continue over a new fiscal package to avoid a further significant degradation of fiscal support. The White House recently indicated it would accede to a $1.5 trillion bi-partisan compromise proposed by moderate Republicans, but little visible progress has occurred despite Democratic Party leader Pelosi indicating she would accept a scaled-back plan to cover the pre-election period. The coming battle over the newly empty Supreme Court seat threatens to scuttle any serious attempts at a fiscal compromise.

Canadian fiscal

September 23 will bring Canada’s Throne Speech, and with it a roadmap for future fiscal policy. With a new finance minister and rumblings of a bold new expansionist vision, there is the possibility of a significant expansion of government, whether in the form of expanded pharmacare, expanded daycare support or a permanent increase in financial support for unemployed Canadians and/or low-income households. The current left-leaning minority formulation of the country’s parliament makes such programs conceivable.

However, with a second wave of the virus now washing ashore in Canada, the focus is reported to have lately shifted back toward efforts to control the virus and measures to sustain the economy through this difficult period. In other words, an emphasis on temporary rather than permanent measures.

The pandemic itself does not demand a permanent expansion of government or a permanent new revenue tool, and it would be most unusual to raise taxes while emerging from the worst recession in a lifetime.

The CERB, Canada’s main emergency support program for unemployed workers, is set to expire on September 27, to be replaced by several new programs. Of the four million people currently in the CERB, three million are expected to transition to a simplified form of employment insurance, with the other million set to be spread across the new Canada Recovery Benefit, the Canada Recovery Sickness Benefit and the Canada Recovery Caregiving Benefit. The EI program has been enhanced both from an eligibility and financial perspective, but ultimately both the EI and the CRB programs will provide slightly less money to unemployed people: a drop from $500 per week to $400 per week.

The CD Howe Institute indicates that those who lost part-time jobs will continue to see their income replaced at as much as 200% of their prior level – so a significant disincentive to work remains. However, those who had previously worked full-time at the minimum wage will now see their wages replaced by just 65% to 86%, depending on the province.

U.S. election update

The U.S. election is now just six weeks away. Polls continue to put Democrat Biden moderately ahead, with an average 7 percentage point lead.

Perhaps more instructive is the probability that Biden will capture the presidency versus the incumbent Trump. The PredictIt betting market indicates that Biden has a 13 percentage point probability advantage. This is significant, but hardly conclusive (see next chart). Furthermore, the race has narrowed somewhat over the past few months, although Trump failed to replicate the big convention bump he enjoyed four years ago.

Biden leads Trump, but gap has narrowed

Note: As of 09/17/2020. Based on prediction markets data and RBC GAM calculations. Source: PredictIt, RBC GAM.

Interestingly, there is significant variation in the assigned probability of a Biden win, depending on the source. The so-called Superforecasters at the Good Judgement Open give Biden a 30 percentage point advantage. Modelling using state-by-state polls yields a 54 percentage point advantage according to fivethirtyeight.com, and a whopping 72 percentage point advantage according to the Economist Magazine (to be clear, this means an 86% chance of a Biden victory, versus 14% for Trump). As such, the election can be described as anything from fairly close to sure-thing depending on the source.

One point of particular uncertainty is the fact that the 2016 election outcome ran contrary to expectations. In fact, New York Times analysis finds that, were the polls as wrong in 2020 as they were in 2016, Trump would win a second term. This is certainly possible. However, let us recognize that it is far from likely, as per the probabilities above. Polling is as much art as science. Different demographics respond to polls at different rates, and with different levels of candor. As such, pollsters have always had to adjust their results to remove these biases. The adjustments prior to the 2016 election were clearly insufficient, perhaps failing to capture the significant number of “shy” Trump voters who were reluctant to reveal their intention to vote for Trump. Those lessons should – at least in theory – now be reflected in the 2020 polls.

Of course, conditions could change over the next month and a half: there are an unusual number of X-factors this time, with more potentially on their way. These include:

• ongoing protests over racial injustice
• a Supreme Court nomination battle
• the prospect of unprecedented mail-in balloting
• the general uncertainty of holding an election during an economic and health crisis.

Congress continues to look as though it may also move into Democrat hands, though the Senate race remains particularly close (see next chart, and note that the two different axes).

A probable Democratic Sweep

Note: As of 09/17/2020. Differential measured as difference between the probability of winning the White House/Congress by the Democratic Party and the Republican Party. Based on prediction markets data and RBC GAM calculations. Source: PredictIt, RBC GAM

U.S. dollar implications

The U.S. dollar has been fairly steady over the past few months. Of greater relevance, it is down significantly from its early March high, and also from the norm of the past several years. Our view is that the dollar is now in a structural bear market. It is set to depreciate somewhat further over the next several years, albeit not as aggressively as the first leg downward.

By extension, other major currencies have been mostly stronger. For its part, the Canadian dollar has picked up 6 cents since its late March low, though this only brings it back in line with the average of 2019. We look for further Canadian dollar strength over the next several years, though not to the extent of most other major currencies.

The purpose of this note is not so much to provide the justification for this currency view: this is grounded in a variety of considerations, including the overvalued nature of the greenback, the loss of U.S. global leadership, the prospect of an extremely dovish Fed, the loss of fiscal restraint in the U.S., and the de-dollarization of global foreign exchange reserves.

Instead, the purpose is to begin the process of highlighting the various implications of this currency shift. Salient considerations include:

• Elevated gold prices are at least in part because the dollar’s status has diminished, leaving other currencies (and the oldest quasi-currency of all – gold) set to potentially benefit.
• A depreciating currency can be moderately inflationary for a country, and modestly deflationary for others. This nudges the U.S. inflation outlook a little higher and the rest-of-world inflation outlook a bit lower.
• A depreciating currency boosts the competitiveness of a country, making the U.S. somewhat more competitive relative to its peers and rendering other countries a bit less competitive. As such, one might expect the U.S. economy to grow incrementally more quickly and others a bit less quickly.
• However, there is another consideration: given the centrality of the U.S. dollar in international transactions, a 1% decline in the value of the dollar is theoretically associated with a 0.6% increase in the volume of trade in the rest of the world. This may offset the aforementioned competitiveness hit.
• Of course, a weaker currency also makes the U.S. technically poorer – less able to buy foreign assets, and with the rest of the world more easily able to afford U.S. assets.
• A weaker U.S. dollar can be a positive for emerging market countries in two main ways. First, they don’t suffer the outflow of capital that frequently arises when the dollar is strengthening. Second, their U.S. dollar debt becomes cheaper in local currency terms.
• Lastly, a weakening greenback means that U.S. financial market returns should be proportionately weaker for non-U.S. investors in local currency terms. The currency decline eats a bit of the return.

So far, the currency move has been only moderate, albeit fast. But if this dollar cycle plays out like many have in the past – representing a sizeable move over the span of perhaps five to eight years – some of these factors could start to become more relevant for the economic forecasting process, and also for financial market returns.

At this juncture, it doesn’t appear that the euro or the Chinese renminbi can reasonably aspire to become the world’s new reserve currency – so no cataclysmic change is underway.

As to whether the U.S. dollar move is a positive or negative, that depends on which of the above variables are most relevant to the party in question. But, as a general statement, it is arguably healthy that the dollar may be on a course to reduce the extent of its overvaluation. Having currencies close to their fair value is arguably optimal, all else equal.

Lasting repercussions

In the early weeks of the pandemic, we engaged in intense brainstorming about the long-term implications of COVID-19. Although six months have since passed, the set of factors hasn’t actually changed all that substantially.

It is nevertheless worth occasionally revisiting these thoughts given their high importance. We have also refined the list in that the items are now sorted into three separate buckets (see table).

Lasting implications of COVID-19

Source: RBC GAM

The first bucket represents genuinely permanent changes. There aren’t all that many. A prominent one is that public debt loads will likely be permanently higher. We simply don’t see a viable path toward paying down the debt that has been accumulated in the effort to withstand COVID-19. For that very reason, interest rates will likely have to be permanently lower than before. This is not to say they can never rise – they should, over time – but that the level of rates will probably be incrementally lower than otherwise at all times along the way.

The second bucket also contains permanent changes, but ones that we deem to have been inevitable even without COVID-19, albeit arriving at a later date. Prominent examples include the leap forward in online shopping (which has roughly doubled), additional automation and more working from home. In fairness, one might expect each of these to retreat a bit in the coming years, but much of the increase will stick. A recent Brookings Institute survey found that one-fifth of office work is expected to shift permanently from offices to homes. These outcomes might have taken five, 10 or 20 years to happen without the pandemic.

The third bucket contains temporary changes, but changes that are nevertheless likely to endure for multiple years, making them entirely relevant to investors. Examples include:

• economic weakness that should persist for a few years
• our expectation that governments will be larger for several years, but ultimately return back to normal
• a preference for low-density living that is entirely understandable today but will probably fade alongside the virus in future years
• the possibility of a temporarily lower steady-state return on investment in response to a lower level of economic activity.

But each of these – and the many other items in this bucket – should eventually unwind.

Supporting the thesis that much will ultimately prove temporary, China – which has nearly eradicated the virus – now reports packed gyms, shopping malls and restaurants, and even movie theatres are now back to 90% of normal.

A San Francisco Fed paper recently concluded that, after past pandemics, the real rate of return declined for about 20 years and did not return to normal for 40 years. This was seemingly because the population was decimated and so capital could not generate as large of a return. However, this pandemic arguably differs in that the loss of human life – while large and tragic – does not come close to matching those prior pandemics. In turn, it doesn’t appear that the rate of return has be lower for decades to come for population-related reasons.

-With contributions from Vivien Lee and Kiki Oyerinde