I’m supposed to be sipping a piña colada while poolside in Mexico this week, but events have conspired to keep me where fruity beverages and warmth remain scarce commodities. The world is in a state of flux right now, and apparently economists are in high demand. No one is more surprised than I!
On the home front, cabin fever has already set in with three weeks still to go before schools re-open. As we scramble for ways to keep the kids stimulated, an intriguing pursuit so far has been a computer puzzle game called Foldit that lets you manipulate proteins in an effort to come up with a design that might neutralize Covid-19 in real life. It could make for an all-time great “What I did over spring break” essay!
The global trend
Covid-19 continues its rapid spread, with a record 10,982 new cases and 438 additional deaths on Sunday March 15 alone. The non-China daily growth rate remains approximately 20% – a worrying pace given the ever-rising base (see chart).
Increasing Covid-19 global cases (ex-China)
Note: as of 3/15/2020. Source: WHO, RBC GAM
The level of official concern has ratcheted higher: the World Health Organization has now officially declared the virus to be a pandemic, and the U.S. and a number of other nations have declared a national state of emergency.
The reach of the virus varies substantially by nation, with a handful of countries having seemingly tamed the disease, whereas others remain very much in its jaws. Still others have only begun their encounter.
On the positive side, China has now almost no new cases, with just 27 confirmed in the latest day (see next chart). This is nearly 150 times lower than in early February. And, to put it into further perspective, there are now roughly 20 times more people who recover from the disease each day in China than are newly infected by it.
The spread of Covid-19 within China is slowing
Note: As of 3/15/2020. Spike on 2/17/2020 due to change in reporting methodology. Source: WHO, RBC GAM
South Korea has similarly tamed its growth rate, though it is not as far along as China, recording nearly 11 times fewer new daily cases than its peak date at the end of February (see next chart).
Spread of Covid-19 in South Korea
Note: As of 3/15/2020. Source: WHO, Macrobond, RBC GAM
Japan has become blurrier, experiencing a roughly constant number of new cases per day since early March. Still, this is a small victory in that it represents a declining percentage growth rate and argues that the transmission rate is therefore gradually declining, if less impressively than in the aforementioned countries.
At the opposite extreme, the rest of the world is now responsible for roughly half of all cases for the first time. This fraction will only grow from here.
Italy is now the most worrying country in the world, in part because its caseload is spiking – 3,497 new cases and 173 deaths in the latest day alone; in part because it now has by far the most cumulative cases (21,157) and deaths (1,441) outside of China; and in part because it is a developed nation and so theoretically says something about the developed world’s ability to handle this outbreak (see next chart).
Spread of Covid-19 in Italy (daily change)
Note: As of 3/15/2020. Source: WHO, Macrobond, RBC GAM
More generally, the locus of concern has now shifted away from Asia and toward Europe. Over half of the world’s new cases are coming from Europe’s major nations, with Spain, France and Germany following on Italy’s heels. Spain is particularly concerning: while it has fewer than a quarter as many overall cases as Italy, it is growing roughly twice as quickly on a percent basis (1,522 new daily cases).
Iran remains another hot spot, with 1,365 new cases in the latest day (12,729 total). Meanwhile, the U.S. data remains highly volatile and somewhat suspect – they claim no new cases in the latest day but 414 the day before that (1,678 total). Canada clicks along with 68 new cases (244 total).
Looking for patterns
Now that China and South Korea have moved beyond their outbreak’s peak, the game is afoot to identify some underlying mathematical constant that might help predict the future trajectory for everyone else. The next chart shows how the number of cases per country has evolved on a population-adjusted basis over time. China and South Korea managed to stabilize their outbreaks after around 20 days of growth (day one is the point at which they reached one case per million population). This is a promising observation, as Italy is now nearing that duration, with a host of others following on its heels. If this pattern holds, the world could become much less nervous by the end of the coming week.
Trajectory of spread of Covid-19 in different countries
Note: As of 3/15/2020. Number of cases per 1 million population in log scale. U.S. first exceeds 1 case per 1 million population on March 4, 2020. Source: WHO, Macrobond, RBC GAM.
However, even setting aside the fact that the definition of the “starting point” is debatable and it is unclear whether the caseload should be examined in absolute or per capita, Covid-19 doesn’t simply run out of steam after 20 days all by itself.
Rather, China and South Korea did some very specific things with regard to social distancing and screening that permitted their trends to flatten out. We discuss these best practices in detail later. The good news is that other countries are finally learning China’s lessons and implementing their own quarantines and testing programs.
The real mathematical relationship probably has more to do with when countries began their own quarantine programs. The most affected Chinese province of Hubei, for instance, started on Jan 23 and its caseload peaked on February 5 – meaning the daily number of new cases continued to set records for a further 13 days before beginning to ebb. That’s a bit longer than we would have imagined given what is known about the incubation period of Covid-19, but not impossible given delayed diagnosis and imperfect quarantining. If this pattern holds, Italy’s serious quarantining began on March 9, with the implication that the country’s caseload could peak on March 22 (next Sunday). The country could conceivably peak earlier than this to the extent that it had a limited quarantine in place beforehand, but on the other hand its testing regimen likely still isn’t up to Chinese or Korean standards, nor are its border controls as stringent.
Optimistically looking even further out, to what the decline trajectory looks like, Chinese provinces appeared to take a further 15 days on average after their daily peak for the caseload to return to virtually zero (see next chart).
Mapping coronavirus in the most affected provinces in China
Note: As of 3/15/2020. New clinically diagnosed cases included for Hubei starting 2/14/2020. Provincial data not available until 2/1/2020.
Lower fatality rate
The true fatality rate of Covid-19 is not yet clear. Calculated simply, 3.8% of infected people have died. There is a raging debate as to whether this constitutes an unreasonably optimistic or pessimistic figure.
On the pessimistic side, this initial estimate could underestimate the danger in that many have only recently been infected and will eventually die after what is often a multi-week period of illness, and furthermore the ability of health professionals to manage the outbreak diminishes as the caseload grows.
On the optimistic side:
- the mildness of the symptoms in many cases means that there are almost certainly many more infected than is officially acknowledged (meaning that the number of deaths should be divided over a larger base)
- one would think that medical best practices should improve rather than deteriorate over time
- there are many examples of jurisdictions managing to keep their fatality rates below 1%.
The Diamond Princess cruise ship remains particularly illuminating because everyone aboard was tested and received proper medical treatment. Even though the average age was fairly old, the fatality rate on the ship was just 1.0%.
A further argument in favour of a lower fatality rate is that the coronavirus is known to be prone to genetic mutations. While this could theoretically result in a drift in an undesirable direction, it is more likely that it becomes less dangerous. The virus is already unusually deadly and unusually transmittable by the standards of coronaviruses. All else equal, it is thus more likely to shift back toward more familiar parameters than to become even more atypical. Additionally, it is the nature of most viruses to become less virulent over time because the most deadly ones kill their hosts too quickly to be reliably transmitted, and with this disease the most ill people are also being isolated and treated, whereas milder strains might be unrecognized and left to run free.
We are more persuaded by the optimistic arguments, and so work with the assumption that the fatality rate will eventually alight in the 0.5% to 1.0% range. This is five to 10 times worse than the flu, but simultaneously five times better than the current raw fatality rate.
Also note that our 0.5% to 1.0% fatality estimate doesn’t mean that this fraction of the world dies. It may still be possible to control the spread of the virus, and even if not, epidemiologists assume that “only” 30% to 70% of the population would contract the virus, resulting in the loss of perhaps 0.25% to 0.5% of the world’s population. Not a happy thought, but an important further qualifier.
Transmission rate now stalled
We have constructed a rough proxy for the number of people each sick person infects. In the early going, this was around 6 in China, and has since fallen well below the critical threshold of 1 that means the disease is in retreat.
For the rest of the world, the transmission rate was also as high as 6, and has since declined significantly. However, for the past week it has been stuck in the vicinity of 3 and even edged slightly higher (see next chart). This is not good, as it means that the average infected person is still passing the disease on to three others.
Fortunately, as quarantining and testing increase outside of China, this transmission rate seems likely to begin falling again, at least within the countries that are now taking it seriously and introducing aggressive social distancing policies.
Transmission rate suggests continued decline in China and growth ex-China
Note: As if 3/15/2020. Transmission rate calculated as 7-day % change of underlying 5-day moving average of daily new infections.
Secrets of success
The key to controlling the spread of the virus appears to be some mix of social distancing, screening and travel restrictions. Furthermore, it is helpful when nations have universal health care, provide a coordinated public health response and communicate well with the public.
Interestingly, not everyone has done it the same way. China has gone hard on all fronts, implementing aggressive quarantines with police enforcement, conducting a remarkable number of tests and tracking cases back to their source, and limiting international travel.
South Korea, on the other hand, has not quarantined to the same extent, but has been particularly vigilant about testing its population and tracing cases. Singapore, another (small) success story, has managed to keep open its schools, instead focusing on closing its borders.
The point is that not every country has succeeded in the same way. However, as the caseload grows, it is unlikely to be sufficient to simply close borders, though that may be a necessary condition. Similarly, the virus could soon become too widespread to make comprehensive tracking and testing feasible, though this will undoubtedly remain helpful where possible. Quarantining seems like the best bet, with evidence already mounting that going half way doesn’t work.
Quarantining has the added advantage that, in the event that it were too late to stop the spread of the disease (a far from resolved idea), such measures would at least “flatten the curve” – make it so that people are infected over an elongated period of time such that hospitals are not so overloaded and can deliver adequate care to everyone.
When Covid-19 first presented itself, the initial instinct was to compare it to SARS. But this comparison has increasingly proven flimsy as Covid-19 now easily exceeds SARS in almost every way. We are therefore in need of more useful comparisons. Three present themselves: the Spanish Flu of 1918-1919, the Asia Flu of 1957 and the Hong Kong Flu of 1968-1969.
The Spanish Flu of 1918-1919 was easily the most serious of the three, ultimately felling around 40 million people worldwide (2% of the world’s population). It was the second most deadly outbreak in recorded history, after the approximately 60 million killed by the Black Death between 1348 and 1351. Roughly 675,000 Americans and 50,000 Canadians died. Notably, the Spanish Flu had three separate peaks over the span of a year, highlighting that such viruses can occasionally re-emerge.
The historical record shows that American cities that failed to quarantine aggressively and in a timely fashion suffered much worse outbreaks than other parts of the country. For sports fans bemoaning the loss of the NHL, NBA and MLB, there is a parallel in history: the Spanish Flu prompted the cancellation of the 1919 Stanley Cup finals.
From an economic standpoint, 1918-1919 experienced a mild global recession. This is not often attributed to the virus, and of course World War I was a further complicating factor at the time, but one wonders if the virus played a bigger role than conventionally imagined. The stock market response to the virus was surprisingly muted, with a peak to trough 11% decline in the Dow, followed by a rally into the end of 1919 that more than undid the damage.
We’d like to think that Covid-19 won’t quite match the Spanish Flu, but it could well resemble the Asian Flu or Hong Kong Flu. The former killed 1.5 to 4 million people globally (70,000 in the U.S.) while the latter felled 1 to 4 million people globally (33,000 in the U.S.). In both cases, there was a global economic recession that occurred at the same time or soon thereafter. As with the Spanish Flu, economic historians have tended to place the blame elsewhere, but it is at least a startling coincidence. U.S. stocks were again down during the worst of the experience, by 21% with the Asian Flu and by 13% with the Hong Kong Flu. In the former case, they rebounded briskly afterward; in the latter case, a full rebound was not forthcoming as the 1970s proved a grim period for equities for a host of unrelated reasons.
While one might debate whether or not these viral outbreaks created the accompanying recessions, what is crystal clear is that none of the events had a lasting effect on the economy. Indeed, one struggles to find them even mentioned in economic textbooks. This is a promising thought with regard to the long-term economic outlook.
We continue to revise our economic assumptions as more information becomes available. To the extent that the virus has continued to spread, quarantining has jumped in the developed world, and China is now reporting serious economic damage, we have been obliged to downgrade our forecasts.
Rather than fixate upon a single forecast, we prefer to continue using multiple scenarios. Simply put, there are many things we don’t yet know, including:
- The extent to which politicians prioritize human life versus their economies. They will likely continue to prioritize the former, but this is not as absolute as it seems. Many countries are still reluctant to engage in widespread quarantining. How this thinking evolves will be important.
- Whether Italy follows the mathematical pattern discussed earlier, recording a peak in new infections in roughly a week’s time. If not, our understanding of the situation may be flawed.
- How much will social distancing reduce the supply of products into the economy? It is not in the least clear what fraction of people will be unable to work, and how much business output will be lost.
- How much will social distancing limit the demand for products, and to what extent will businesses engage in a more enduring retreat via layoffs and cancelled capital expenditures? To what extent will households do the same via less home buying and fewer car purchases?
- To what extent will a temporary hole in the economy turn into something more enduring/deeper as companies encounter liquidity and perhaps even solvency issues?
Our Positive, Medium and Negative scenarios encapsulate the range of likely outcomes (see next chart).
Covid-19 economic channels and likely effect
As at 03/15/2020. Source: RBC GAM
The economic implication for each scenario has been further downgraded relative to a week ago. We now believe the most likely is the Medium scenario rather than straddling the Positive and Medium scenarios as we were last week.
The assumptions for each economic channel are laid out in the accompanying graphic, with the end result that a Medium scenario would have developed economies underperform their normal rate of ascent by 1.0 to 2.0 percentage points in 2020. This is a significant hit, arguing that Eurozone GDP is likely to decline outright over the year. The U.S. should still manage some slight growth in 2020, but Canada may go roughly sideways due to the additional damage of the recent oil shock on the economy. But let the record show that even the U.S. and Canada are now more likely than not to suffer technical recessions as their economies decline for two consecutive quarters.
To be sure, it is still possible that the economic hit might only be -0.25 to -1.0ppt should the disease somehow fail to gain traction in North America and calm somewhat in Europe. Perhaps a more effective treatment will be found, or even a vaccine.
Conversely, if the experience ultimately replicates the death toll and persistence of the Spanish Flu, the damage could well be -2.0 to -4.0ppt from GDP. Should the genie prove difficult to put back in the bottle, the economic hit could last a year or longer. From a calibration perspective, this squares with a World Bank report from the late 2000s that estimated a 3.0ppt hit to the global economy should a similar event occur in the future. While that report failed to imagine quarantining on a large scale, it is equally the case that health standards, medical technology and the ability to work from home have improved massively over the past century.
Based on reasonable assumptions about the relative size of the economic hit to different sectors, we arrive at a stylized sector-level forecast for U.S. GDP as depicted in the next chart.
Medium scenario: U.S. 2020 GDP growth deviation from normal
Note: As of 3/16/2020. RBC GAM estimates of GDP growth deviation from normal. Source: BEA, Haver Analytics, RBC GAM
It is still too early to observe significant economic damage in the developed world, with U.S. weekly jobless claims remaining surprising low even as the level of tension ratcheted higher in recent weeks.
But the economic pain is already becoming visible in China, where we now know that Chinese car sales are down 79% from a year earlier in February, with retail sales more generally down 20%. Fixed asset investments are down 24% and the country’s unemployment rate has increased by a full percentage point between December and February. Even as evidence mounts that the country is beginning to restart its economy, we feel compelled to downgrade our growth forecast for China yet again. This is in part because of these latest data points, in part because China will be hit by a boomerang of weakness from the rest of the world, and in part because the country doesn’t appear to be concealing its economic damage as much as we had initially assumed. As such, our China 2020 growth forecast pivots from 5.0% to just 4.0%.
With a Medium scenario forecast that now includes multiple quarters without growth for the developed world, a technical recession seems more likely than not. To be clear, this is not a certainty, but we view it as a greater than 50% likelihood.
Alas, we cannot rely upon our yield curve-based recession model to render this verdict, as the model does best when the recession is some distance away rather than staring us in the face. Once central banks are locked into aggressive easing, the yield curve unavoidably steepens and renders the model voiceless.
Similarly, our business cycle scorecard doesn’t have a great deal to say. This recession – should it arrive – is entirely the result of an exogenous shock rather than the natural ebb and flow of the business cycle. It would be fair to argue that the lateness of the business cycle means the economy may be more vulnerable than usual to a drawn-out multi-faceted recession, rather than the short, mechanical supply-side recession one would normally associate with a pandemic.
Fortunately, policymakers are now busily at work, seeking to minimize the extent of any downturn. More on that shortly.
While supply chain issues could theoretically prove inflationary, the great bulk of the other forces at play should be deflationary. Add in the impact of sharply lower oil prices and inflation should be unusually low in 2020.
Central banks cut
Central banks have been extraordinarily busy delivering extraordinary stimulus. The U.S. Federal Reserve has now announced two inter-meeting cuts, slicing its policy rate down to an all-time low of 0%. The Bank of Canada has also cut by 50bps twice – once inter-meeting – to reach a 0.75% policy rate. Other central banks have been similarly busy, with the Bank of England easing and the ECB lowering the rate at which it lends funds (though not its benchmark deposit rate). The People’s Bank of China has been actively easing throughout.
Cumulatively, we are working with the assumption that developed-world central banks will deliver an average of 100bps of monetary stimulus in response to Covid-19. The Fed has already done more than this, but others such as the ECB and the Bank of Japan are more limited given their starting point. We do not expect additional central banks to embrace negative interest rates. Most continue to believe that 0% is as low as one should go.
Quantitative easing has so far only played a small role, with the ECB moderately increasing the size of its ongoing bond purchases. The Fed will continue to inject $60 billion per month into the bond market, pivoting away from its prior short-term focus to support a broader swath of the yield curve. To the extent central banks determine that more economic assistance is needed, there remains ample room to deliver more QE, particularly in North America. This is the main remaining upside risk from monetary policy.
A valid criticism is that rate cuts are like pushing on a string at a time when workers are being told to stay home, but it nevertheless makes outstanding debt cheaper, increases the money supply and serves as a theoretical confidence boost (though, so far, this last part has not yet engaged).
Arguably more important than the rate cuts themselves are the many liquidity measures central banks are introducing to keep the banking sector moving, and to keep credit in the real economy. These actions already amount to trillions of dollars of support, though they are inherently temporary and so do not carry the same baggage as quantitative easing. Central banks will do more as needed, keeping illiquidity at bay.
As a further key effort to keep banks in a position to lend, several countries have now reduced the capital buffer that their financial institutions are obligated to hold, with the effective result that banks are suddenly capable of lending much more than before. The very purpose of these buffers were to leave room for an emergency need such as this.
Fiscal policymakers have also been busy on a number of fronts. The most important thus far are non-monetary in nature. Key actions include:
- Imposing quarantines and travel bans to limit the spread of the disease.
- Setting up special health care programs to provide large-scale testing for Covid-19, and to handle the anticipated surge in cases.
- Declaring a state of emergency to allow expedited government decision-making and to unlock special funds.
- Some countries have also made other targeted policy choices. Italy has delayed mortgage payments and certain other loan payments for a temporary period of time; the U.S. is halting interest payments on student debt.
Of course, more traditional fiscal mechanisms are also being pursued. Many countries are pursuing policies that will help the economic sectors and especially small businesses set to be most hurt by a temporary economic stoppage.
Demonstrating the extent to which countries are shifting their thinking, Germany has explicitly abandoned its longstanding obsession with maintaining a surplus, the EU has emphasized that its countries are permitted to run significant deficits during emergencies, and U.S. Democrats and Republicans have already managed to pass two pieces of legislation unleashing funding for Covid-19.
So far, the sums involved have tended to be fairly small – a few billion here and there – but we suspect the fiscal actions will become bolder with time as the focus shifts from preparing the health care system to supporting the broader economy, with additional global fiscal stimulus worth 1ppt of GDP eventually unleashed. For context, that would amount to around $200 billion in additional spending/tax cuts from the U.S. alone.
Combining monetary stimulus and fiscal stimulus, policymakers may manage to deliver an economic boost worth around 1.5ppt of global GDP. This is theoretically large enough to plug the great bulk of the hole created by the virus. However, the virus hits growth very quickly, whereas stimulus engages with a lag. As such, the economy is still likely to descend into a recession, but the subsequent recovery could be expedited.
This has been a very challenging period for financial markets. Risk assets are sharply lower and volatility remains exceptionally high. The S&P 500 has now entered bear market territory (-20%) for the first time in 11 years, and a large fraction of the world’s markets have joined it in that retreat. For their part, government bond yields are setting all-time lows, providing a helpful offset in the portfolios of balanced investors.
There is very little clarity in the short-term market outlook. In fairness, the way forward is rarely clear. But Covid-19 presents many special challenges, including the extent to which recent quarantine efforts will begin to work and how extensively and enduringly the economy will be damaged. The very fact that we have opted to work with economic scenarios rather than a single comfortable forecast speaks to this.
We can nevertheless make a few hopefully useful observations about the longer-term outlook.
Bond yields are now exceptionally low. It is not that these are unjustified – after all, central banks are again pressing on the accelerator – but rather that it will be very difficult for the fixed income market to generate attractive returns from this low starting point.
On the stock market side, while the average bear market decline is 36%, bear markets within a broader supercycle bull market (which we believe is underway) have averaged a drop of “just” -27%. That’s in the realm of where equities have already descended. Furthermore, periods of high volatility have historically represented an attractive time to buy risk assets.
It remains extremely instructive that Chinese equities have been among the strongest performers in the world, despite the many Covid-19 cases and fatalities in the country. The market appears to reward bold actions to control the disease, even if that action does short-term economic harm. As a result, investor focus should arguably be on the extent to which each country pushes to control Covid-19 as opposed to the temporary economic damage incurred along the way. It is promising that many countries are getting more serious about this, though the challenge is immense.
Even pessimistic assumptions about earnings and valuations point to attractive equity market returns of roughly 5% to 9% over the coming three years. From an economic standpoint, the 1-2ppt hit to developed-world growth should ultimately be unwound by what we expect will be a similarly sized amount of monetary and fiscal stimulus.
History shows that past pandemics – even major ones like the Spanish Flu – imposed only temporary damage to equities, and of a magnitude no worse than the current market drop. Then again, the quarantining and associated lost output could be more notable this time.
In conclusion, it remains unclear how markets will evolve over the coming weeks and months. But history and valuations argue that opportunities exist for those possessing a sufficiently long investment time horizon.
-With contributions from Vivien Lee and Graeme Saunders