The latest week has brought a mix of positive and negative developments regarding the global pandemic. On the positive side, the economic recovery is continuing, the latest North American job numbers were surprisingly good and financial markets are fairly happy.
However, there are significant if incremental negative developments. The number of new daily cases of COVID-19 is growing globally. Several emerging-market countries are particularly affected. And a significant fraction of U.S. states are now suffering a rising caseload.
The number of new daily COVID-19 infections around the world has leapt higher once again, to as many as 130,000 per day. The trend is most certainly a rising one, having spent early April through mid-May at around 80,000 per day (see chart). The cumulative number of infections now adds up to nearly 7 million.
Spread of COVID-19 globally
Note: As of 06/08/2020. Spike on 02/13/2020 due to methodology change. Source: ECDC, Macrobond, RBC GAM
The number of COVID-19-related deaths hit an inauspicious milestone of 400,000 recently, though the trend of daily deaths is mercifully now less than half of that reported in mid-April. We have tended to argue that the death trajectory is more accurate than the infection trajectory. Alas, even the death rate has now flattened out globally – it is no longer significantly declining. In fact, we have replicated our standard transmission rate calculation, swapping out infections for fatalities. The concept is a flawed one – people who have died aren’t passing death on to others in the same way that people who are infected pass on the infection to others. But the underlying math essentially captures whether the virus is likely spreading or retreating. It is now again tentatively the former (see next chart).
Global transmission rate hovering around key threshold of one (using deaths)
Note: As of 06/05/2020. Transmission rate calculated as 7-day % change of underlying 5-day moving average of new daily deaths. Source: ECDC, Macrobond, RBC GAM
Turning back to the more conventional usage of the transmission rate, Sweden now appears to have a serious problem, the U.S. is on a precipice, and the bulk of the developed world is enjoying a transmission rate below one (see next chart).
Transmission rate under one suggests COVID-19 in retreat
Note: As of 06/08/2020. Transmission rate calculated as 7-day % change of underlying 5-day moving average of new daily cases. Source: ECDC, Macrobond, RBC GAM
The U.K. now has the strictest social distancing among major nations, and as its reward the country has now managed to reduce its daily infection count to just over 1,000 from around five times that at its peak.
The Canadian story remains familiar: a mostly declining infection trend, from around 1,500 cases per day at its peak, to roughly 500 today. But Ontario remains stuck with a sideways trend, and is responsible for the bulk of the country’s cases.
The U.S. is going sideways at the national level, with over 20,000 new infections per day. However, the narrative varies massively depending on the state, as discussed later.
New Zealand merits a moment of mention because the country has not recorded a new case of COVID-19 in 17 days, allowing it to now remove all remaining social distancing policies. New Zealand is fortunate in its low population density and isolation from the rest of the world, but nevertheless deserves commendation and is now reaping the rewards of earlier efforts.
Emerging market nations remain a growing hotspot, with a significant increase in the number of infections. The six most affected EM nations are now recording about twice as many new cases per day as the top six developed countries (see next chart).
COVID-19 hitting emerging market countries now
Note: As of 06/05/2020. DM aggregates case count from France, Germany, Italy, Spain, U.K. and U.S., and represents 45.4% of global cases. EM aggregates case count from Brazil, India, Iran, Peru, Russia and Turkey and represents 26.9% of global cases.
Emerging market danger
Even as EM nations become the new epicenter of COVID-19, not all can afford to maintain the economically damaging quarantining that has been in place since March. A Pakistan court decision recently re-opened the Pakistan economy despite a rising caseload. India has also been easing restrictions even as its daily infections rise.
This obviously presents the distinct prospect that the virus will continue to accelerate in some EM nations. The only thing capable of stopping the virus in such countries is a powerful drug, vaccine or herd immunity. None are near-term propositions.
However, several caveats are in order:
- Human behavior is also a key determinant of the virus’ spread. Illustrating this, even before governments decreed that restaurants be closed across the developed world, restaurant reservations had plummeted to near zero. As such, even as EM governments permit economies to restart, individuals and businesses will likely continue to demonstrate considerable caution. For instance, even as Indian rules have eased, actual measures of mobility remain quite low.
- Although they may lack the ability to deliver massive stimulus or comprehensive health care solutions, EM countries are not without options as they reopen. It costs little to stay two metres away from other people, and masks are increasingly available and seemingly effective. Kenya has now mandated that all people wear masks in public, with any violation punishable by imprisonment.
- The poorest developing countries have enormous natural advantages in their relative youth and low obesity levels – some of the strongest predictors of COVID-19 fatality. There are also some claims that vaccination against tuberculosis – which is widespread in developing nations – may confer some cross-immunity to COVID-19. A number of researchers also propose that warmer weather may limit the spread of COVID-19 – a positive development for the many tropical and sub-tropical poor countries.
- COVID-19 remains a smaller problem than malaria for many African nations. A quarter million people have died from malaria since the start of 2020, predominantly in Africa – far more than have died from COVID-19 in Africa. The parallels are imperfect. But to the extent these countries have found a way to continue functioning while plagued by malaria, they may prove adept at doing the same with COVID-19.
- Some EM countries have actually enjoyed spectacular success controlling COVID-19, including Vietnam. China, of course, has managed to mostly quiet the virus domestically. It is far from universal that EM countries are struggling, let alone set for an explosion of new cases.
- The expanding presence of the virus doesn’t necessarily predict economic doom. To the extent the bulk of the economic damage comes from government orders to remain home and for businesses to shutter, the acceleration of the virus could even be associated with a revival of the economy. Both are happening for the same reason: the easing of restrictions. Of course, if the virus becomes sufficiently problematic, such countries might have to shut down anew. But the threshold would presumably be quite high.
Divergent U.S. states
The many U.S. states have pursued different approaches to dealing with COVID-19, and so have had quite a varied experience with the virus. Worryingly, quite a large fraction now report a transmission rate that is greater than one (see next chart).
Transmission rate, U.S. states
Note: As of 06/07/2020. Transmission rate calculated as 7-day change of underlying 5-day moving average of new daily cases. States below dotted line at one have decreasing new daily cases. Includes D.C. Source: The COVID Tracking Project, Macrobond, RBC GAM
Indeed, we now count 22 states with a rising case trend (see next chart). These include Florida, Texas, North Carolina and South Carolina. California is on the cusp. Given the large number of jurisdictions, one would expect random noise to push a few into positive territory at any one time -- but not this many, and not so persistently. Conversely, large states with particularly prominent improvements include Illinois, Massachusetts, New York, New Jersey and Ohio.
Number of U.S. states with transmission rate above key threshold of one
Note: As of 06/07/2020. Transmission rate calculated as 7-day change of underlying 5-day moving average of new daily cases. Transmission rate above one suggests increasing new daily cases. Includes D.C. Source: The COVID Tracking Project, Macrobond, RBC GAM
In past weeks, we tended to view Georgia as the canary in the coal mine. It’s a state that had been particularly eager to reopen its economy, and thus at particular risk of suffering a rising infection rate. However, this theory has just been blown out of the water. The latest mobility data for Georgia reveals a sharp downward revision that has taken Georgia from among the most aggressive to reopen their economies to a much more muted position. This may help to explain why Georgia hasn’t actually suffered a spike in infections over the past month.
Interesting, there would appear to be a red/blue political divide in terms of COVID-19. Initially, it was very much the “blue” coastal states that suffered the worst COVID-19 outbreaks, presumably because they have many high-density big cities and host many international travelers. More recently, however, the pattern has begun to reverse (with some exceptions). The blue states have tended to enjoy a declining caseload after implementing especially aggressive quarantines.
In contrast, the “red” states that avoided the initial outbreak have tended to implement lighter policies and be keener to restart their economies. As such, it is the red states that now appear to be suffering a rising caseload. To be sure, there are prominent exceptions on both sides, with California the most obvious one. Of course, recent anti-racism protests in large cities could yet reverse that pattern yet again given the close proximity of the protestors.
The main issue is that, with both EM nations and a significant minority of U.S. states, we are no longer talking about if the virus revives, but rather what these jurisdictions will do about it now that it is happening. The virus math is not friendly. The pattern tends toward exponential growth fairly quickly.
Perhaps the residents of these places will adjust their own behavior, meaning that governments don’t have to reverse their rules.
But if this isn’t enough, will policymakers be willing to shut down their economies again? Politicians will be extremely reluctant to admit to their past mistakes, and even more so when it simultaneously hurts their economies. In the U.S., the pandemic has been politicized to the extent that this is a further constraint on politicians reversing course. There may be a temptation to let the virus run, especially to the extent health care capacity has increased since March.
To add to this uncertainty, it is furthermore unclear how we should be viewing such troubled jurisdictions. Are they places with the greatest potential for economic growth to the extent quarantines have largely been lifted there? Or are they instead the places with the worst economic prospects to the extent they may have to quarantine all over again.
Economic revival risks
We have been hard at work developing a methodology to determine which jurisdictions are at the greatest risk of requiring a further round of economic lockdown. Relevant variables include:
- The transmission rate: whether the virus is actively expanding or retreating.
- The rate of daily infections per capita: how widespread the virus is within a jurisdiction.
- The level of social distancing: how hard jurisdictions are working to control the virus.
- The decline in social distancing since the peak effort: how much jurisdictions have returned to normal.
Using a scorecard approach, we combined these inputs to assess which countries are at most risk (see next chart). The work has not yet been applied to U.S. states.
Likelihood of continued spread based on lockdown severity and current transmission rates of COVID-19
Note: As of 05/29/2020. Likelihood factors in lockdown severity and amount restrictions eased, transmission rate, and population-adjusted prevalence of COVID-19 indexed to max of featured countries. Source: Google, University of Oxford, Apple, ECDC, UN, Macrobond, RBC GAM
Some findings are surprising. France and South Korea are near the top of the risk list, though largely because they have had mini-recurrences of the virus that we suspect will prove short-lived. Similarly, it is surprising that India is near the bottom, but despite its case growth it has very few cases relative to its population and its quarantines are still stricter than most.
Other countries land largely as expected. Sweden and the U.S. are near the top of the risk chart, with Japan near the bottom. Canada and the U.K. are roughly in the middle.
The latest economic data reveals a variety of interesting things.
Our real-time metrics mostly continue to point to a reviving economy, though hours worked by hourly workers in the U.S. argue that the pace of the recovery has likely slowed. The hours-worked data is only marginally higher than the week before (see next chart).
Percentage change of hours worked by hourly workers in the U.S.
Note: As of 06/06/2020. Impact compares hours worked in a day vs. median for the same day of the week in January 2020. Source: Homebase, Macrobond, RBC GAM
Another non-traditional indicator is the U.S. hotel occupancy rate, which has rebounded from a low of 21% to 32%. Perhaps the most surprising discovery in all of this isn’t the welcome rebound, but rather the fact that the occupancy rate never fell below 21% despite the utter collapse of business travel and tourism.
In Canada, a fascinating piece of journalism on the pawnshop and payday loan industries corroborates tentative evidence that low-income households have frequently ended up earning more money due to generous government transfers. The pawnshop operators reported that the quantity of goods being pawned has declined radically, people have repurchased their previously pawned goods, and the remaining products available for sale have largely been sold to others with money to burn.
Similarly, the payday loan industry in Canada reported a sharp 84% decline in business during the first weeks of the crisis. This would presumably be in significant part because fewer people had paycheques to pledge for a loan and also due to diminished spending opportunities. Conceivably it is also because household incomes have been supplemented by government transfers.
Bridging the gap between unconventional and traditional economic data, Canada’s CFIB (Canadian Federation of Independent Business) Business Barometer is updated every two weeks and conveys the extent to which Canadian small businesses have enjoyed a revival in recent weeks. The measure is around two-thirds of the way back to normal already (see next chart).
Canadian business sentiment tumbled, but has rebounded nicely
Note: CFIB index as of May 2020, GDP as of Q1 2020. Source: CFIB, Statistics Canada, Haver Analytics, RBC GAM
The U.S. payroll data for May constituted a very pleasant surprise, adding 2.2 million jobs. This was in contrast to a consensus forecast for the loss of a further 8 million jobs. It isn’t often that there is a 10 million job surprise when forecasting over the next decade, let alone the next month.
Of course, the return to job creation doesn’t mean that layoffs have stopped. To the contrary, there is still plenty of second-round damage occurring to the economy. A huge 7.7 million unemployed people found jobs, a further 5.4 million people who weren’t even in the official labour market also found jobs, while 9.3 million other workers lost their jobs. Those numbers don’t quite add up to the official payroll survey total because these numbers come from the household survey, which estimated the creation of 3.8 million jobs in May.
The official unemployment rate fell from 14.7% to 13.3%, but the statistical agency indicates that the true unemployment rate actually fell more substantially, from 19.7% to 16.3%. Recall that 21 million jobs were lost in March and April. It will take significantly more job creation before the unemployment rate is normal again.
This return to job creation makes complete sense in that we know economic output rose significantly in May and it would be impossible to produce significantly more adding to the workforce. All the same, the improvement was nevertheless unexpected and remains somewhat mysterious to the extent that continuing jobless claims had continued to rise over the relevant reference period, such that 13 million more people were receiving unemployment benefits over the same time period. This is hard to reconcile. Only subsequently have continuing jobless claims started to decline (see next chart).
U.S. insured unemployment rose slightly in the latest week
Note: As of the week ending May 23, 2020. Source: DOL, Haver Analytics, RBC GAM
In Canada, the job numbers were also surprisingly good in May, rising by 290,000. Almost all of the gains came from Quebec where the decline had been especially steep beforehand. However, in contrast to the U.S., the unemployment rate nevertheless rose, by 0.7ppt to 13.7%. This was mathematically possible because more people rushed back into the labour force (looking for jobs) than actually received new jobs.
Traditional data showing extent of economic decline
As we await more economic data detailing the extent to which the economy has recovered in May, we can nevertheless look back on a variety of charts that convey the depth of economic decline. The first chart shows the decline in U.S. retail sales and industrial production; the second chart conveys the decline in consumer spending pitted against the remarkable spike in household income.
COVID-19 dealt a devastating blow to U.S. economy
Note: As of April 2020. Source: U.S. Census Bureau, Federal Reserve, Macrobond, RBC GAM
U.S. consumer spending fell significantly despite government aids
Note: As of April 2020. Shaded area represents recession. Source: BEA, Macrobond, RBC GAM
The Bank of Canada’s latest decision – and first under new Governor Macklem – stuck largely to script, failing to introduce any significant changes to Canadian monetary policy. To the extent the economy is staging a tentative revival and financial markets are reasonably calm, this is a prudent stance.
Finally, and emphasizing that the initial economic damage in the Eurozone is set to be larger than to the U.S., German industrial production was reported to have dropped a sharp 25% year-over-year in April.
Medium-term GDP outlook
There are many ways to create a GDP forecast. Our initial focus has been on the short-run effects of COVID-19. This makes expenditure-based (consumption + business investment, etc.) and sector-based strategies the natural choice.
But as we think about the medium-term outlook -- call it 2022 to 2025, rather than merely 2020 and 2021 -- another approach is needed. The best strategy is arguably to focus on a different mathematical identity: the fact that the amount of work plus the productivity of the workers must also amount to GDP.
The first of these variables can be simplified into population growth. Globally, COVID-19 could in a worst-case scenario kill something like 0.4% of the world’s population and perhaps marginally reduce the fertility rate for a brief period of time. The actual effect on hours worked should be even smaller than this, as it is disproportionately retired people who are felled. All told, it is unlikely that the effect will be materially greater than a 0.5% diminishment of potential hours worked over the medium run. Furthermore, this is a one-time effect: the population shrinks by that amount once rather than growing more slowly every year by that amount.
In fairness, at the national level, immigration also plays a significant role. Over the short run, this further lowers the size of a country’s population, by as much as 1% in immigrant-dependent places like Canada and less elsewhere. We do not assume that immigration growth will be significantly diminished over the medium or long run, though this is debatable in places like the U.S. that may be further spurred in an isolationist direction.
While a newly configured world undoubtedly provides new opportunities for innovation and profit, the dominant effect is likely to be a reduction in productivity growth. A variety of channels may prove relevant.
In economics, the term hysteresis refers to permanent or quasi-permanent economic destruction that sometimes results from a recession or crisis. In the present situation, it seems reasonable to imagine that some jobs and sectors won’t be returning with any haste. A significant fraction of the world’s passenger airplanes, for instance, will probably go unused for quite some time. However, we note that after the global financial crisis there was a widespread expectation of deep hysteresis and yet ultimately economic activity broadly returned and unemployment rates achieved new modern-day lows.
More generally, productivity growth tends to be diminished for many years after a major shock due to such considerations as an oversized capital stock, diminished wealth and elevated risk aversion.
In the present context, social distancing may result in less collaboration. The retreat of globalization presents a further headwind. For that matter, all it would take is for the average employee now working remotely to suffer a nearly imperceptible 5% decline in their personal productivity for the world to lose a full year’s productivity gain (though some research finds that remote workers are more productive, on average).
Also recall that higher public debt loads will require servicing, perhaps on the order of 0.3% of GDP. This is very different than suggesting that 0.3% of GDP must vanish, as lenders are being paid this money and can recirculate the funds into the economy. Nevertheless, as the research of Reinhart and Rogoff found over a decade ago, higher debt levels do appear to be associated with slower productivity growth.
Overall, it remains a shot in the dark to estimate the extent to which productivity growth might be diminished over the medium run due to COVID-19. For nearly a decade after the global financial crisis, productivity growth was diminished by almost a percentage point. We’re hopeful the lingering damage isn’t quite so large this time to the extent the crisis itself may prove shorter-lived. We are tentatively assuming something like a 0.5% hit to annual growth over 2022 to 2025, consistent with a potential growth rate that might be below 2.0% in the U.S.
It is mathematically obvious that when household spending has cratered and housing income has bloomed (thanks to government stimulus), the residual must appear in a higher personal savings rate. This is a surprisingly common occurrence during recessions – even ones without such generous government stimulus. The instinct to spend less is very strong during recessions as people become risk averse.
The U.S. personal savings rate was already unusually high before COVID-19 arrived, at nearly 8%. For context, it was as low as 2% in 2005. The increase over the subsequent 15 years was attributable to a variety of factors, including:
- less easy credit
- the related end of the housing boom, and
- an aging population that is now in its peak saving years.
But the spike in the savings rate since February is completely unprecedented. We first got a hint of this in the March figures, when the savings rate jumped to 13%. It then blew the lid off of any kind of modern-day precedent with a further leap to 33% in April (see next chart). Those are nearly Chinese levels of saving.
U.S. personal saving rate hit record high amid coronavirus pandemic
Note: As of April 2020. Shaded area represents recession. Source: BEA, Macrobond, RBC GAM
Let us acknowledge that a fair chunk of this savings spike is likely involuntary: it is currently hard to go on vacation right now and people remain shy about visiting stores. The savings rate will obviously settle back down as people become less risk averse and as government stimulus starts to fade over the summer. But it wouldn’t be a surprise if it remained higher than the prior norm for as long as a few years.
As an aside, if you are wondering whom governments and businesses are borrowing from during this crisis, the answer would appear to be households.
A further thought: the higher savings rate suggests not just that governments probably went a bit overboard in terms of the size of the stimulus package, but also that the fiscal multiplier is probably somewhat lower than initially imagined since not all of the additional money is being spent.
For the first time since the arrival of COVID-19, another story has bumped the virus from the front pages: the killing of George Floyd by Minneapolis police and the widespread anti-racism protests that have resulted.
We can conceive of four implications from these events.
- Most relevant in a societal sense is the prospect that racism may decline and police conduct may improve. Given the number of protests, the prominent voices speaking out and the unique lack of distractions for a world still substantially in lockdown, this seems entirely possible. Still, let us not forget that there have been past potential inflection points – the Rodney King beating, trial and riots are particularly prominent – that ultimately failed to substantially change society. Somehow, the prospect seems better this time, though far from certain.
- Turning back to the more mundane economic implications, we do not believe there has been sufficient damage to infrastructure during protests that there will be any palpable effect on economic growth beyond the municipal level.
- The large number of people protesting closely together and without universal mask usage suggest that some parts of the U.S. could suffer an increase in COVID-19 infections over the coming two weeks.
- There may be implications for the election this fall. In general, to the extent the protests represent a widespread awakening, it may tilt voters somewhat to the left – favoring Biden over Trump. However, this is hardly a universal sentiment: it could merely energize each party’s respective base, much as the Trump impeachment failed to tilt the political balance significantly toward the Democrats. Another political consequence is that Democratic nominee Biden is increasingly expected to name a black running-mate. On the aggregate, betting markets are increasingly of the view that Biden has a better chance than Trump at claiming the White House, though it is still very close.
We’ve generally congratulated policymakers for a job well done: their support for economies was aggressive, fast, and fairly well targeted. Important lessons were clearly learned from the financial crisis.
However, the policymaking has been far from perfect. Allow us to grumble about a few things.
Preparedness: Some countries, such as Canada, bragged in advance that their experience with SARS left them especially well prepared for future pandemics. No doubt there were subtle ways in which Canada enjoyed an advantage relative to other countries. But in the end the country’s experience was rather similar to other, theoretically less prepared, countries.
Masks: Representing something of a personal hobby horse, there was arguably insufficient emphasis placed on the importance of wearing masks. In fact, it was actively discouraged in some countries. It is certainly understandable that there was initially a mask shortage and that health care workers had the most urgent claim, but it took the World Health Organization until last week to begin recommending face masks. Meanwhile, the countries with the greatest success controlling COVID-19 were to a disproportionate degree mask wearers.
International borders: Had international borders been closed with greater haste, many countries might have avoided a significant COVID-19 outbreak. But fears of offending other countries prompted sufficient heel-dragging that the virus found its way nearly everywhere.
Late quarantining: As soon as it was clear that a) China was having to fight quite hard to control the virus and b) the virus was beginning to spread internationally, other countries should have implemented obligatory quarantining. However, there was a reluctance to damage economies until it was 100% clear the virus was a domestic problem, by which point it was too late.
Excessive stimulus: This is a provocative claim as many households and businesses are barely getting by even with all of the stimulus delivered, but at a minimum some parties received too much stimulus given that U.S. household incomes spiked in April and roughly a third of all income is now being saved. Whether it was truly possible to better separate those needing help from those who didn’t in real time remains an open question, admittedly. Furthermore, governments may have been squeamish about transferring more money to someone who lost a higher paying job than someone who lost a lower paying job.
Mis-targeted stimulus: It may have been better if household support had come primarily via wage subsidies as opposed to unemployment cheques. The former would have maintained a bond between worker and company, whereas the latter sever that link and make it harder for the economy to revive.
Re-opening too early: As previously discussed, quite a number of U.S. states appear to have re-opened their economies prematurely. This may become a serious problem, and will cause further loss of lives and potentially additional economic damage.
Democracy: As governments rushed to implement emergency stimulus and health care measures, opposition parties and standard legislative procedure have in some cases been shoved to the sidelines. A degree of this is defensible given the unusual circumstances. But in some cases it has gone too far and sets a dangerous precedent.
Tightening regulations: In Canada, CMHC (Canada Mortgage and Housing Corporation) has significantly tightened its mortgage insurance eligibility requirements. This is entirely defensible from the perspective of protecting the crown corporation’s balance sheet and by extension the taxpayer’s money given the possibility of a marked housing slowdown in Canada over the coming years. Furthermore, to the extent Canadian home prices remain high, it may also be desirable in a structural sense. However, CMHC also serves as a vessel of the government, and in this context it is poor timing to be doing something that adversely affects an already suffering sector of the economy.
Closing public spaces: Closing parks and public spaces was probably a mistake, pushing people in need of fresh air and exercise onto narrow sidewalks, and creating pent-up demand that has resulted in subsequent overcrowding. Germany seemingly did this better, having enforcement officers ensuring sufficient social distancing within parks as opposed to keeping people out.
Sweden: Sweden pursued a very different strategy than most other countries, issuing minimal social distancing requirements. However, the country now suffers among the highest infection rates in the world and its chief medical officer has now expressed regret at pursuing this strategy.
To be clear, governments and their agents mostly deserve credit for the force, speed and precision of their efforts. It could have been a lot worse. But it has been far from perfect.
-With contributions from Vivien Lee and Graeme Saunders