The outbreak of the coronavirus in China (COVID-19) is poised to exact a significant toll on global growth. Our revisions incorporating this shock have already lowered our forecast of 1Q20 global GDP growth to the weakest pace of the expansion (Figure 1). An expected plunge in global factory output in February by its largest magnitude in at least half a century (excluding the global financial crisis) is at the center of the downshift. However, recognition that the disruptive effects of the virus are dramatic does not temper our conviction that this impact will prove transitory. By midyear, macroeconomic activity should bounce back to offset most of the earlier shortfall. With the underlying supports for growth intact (recovering sentiment, firming job growth, rising wages, and easy polices), our forecast for a return to above trend growth in 2H20 remains intact. Consequently, gauging the size of the near-term growth dip is primarily important for determining the magnitude of the midyear bounce that we expect to punctuate the return to growth later this year.
This note provides a tracking guide to help navigate through what we believe will be a highly disruptive but ultimately transitory event. Amid the swings, we will be watching for several key features to unfold in the coming weeks and months:
• A reliance on alternative data for China. With China’s official releases delayed by the LNY and January-February readings reported together, we need to look elsewhere to track China activity. As a result, we are focused on an array of daily indicators that we report in our Daily Economic Briefing (link). At present, they show the collapse in China’s activity during the first half of February. These indicators should bottom next week, and then display a solid recovery path into March.
• IP slide is sharp but expectations to hold up. February industry and international trade releases likely will bear the brunt of the shock. We forecast global factory output to fall 3%m/m this month. Less important than the magnitude of the slide is our expectation that it will be heavily concentrated in China. We project global output ex. China (ROW) to drop 0.5%m/m. The global PMI report should also point to ROW resiliency. We expect the ROW PMI output index to remain above 50 this month with the future output index holding above 60, both signaling that China’s disruption is generating limited spillovers.
• Limited spillovers outside global industry. Although the China shock will have broad spillover effects to those countries with close travel and supply chain links, we see limited spillovers outside the region. Indeed, a modest drag from weaker export demand outside Asia should be cushioned by the benefits of lower interest rates and energy prices. To this end, US and Western European indicators of consumer sentiment, labor markets, and credit availability should be largely unaffected by the February disruptions in global industry.
• Limited financial market stress. China’s concentrated shock should stress small and medium-sized business cash flow but policymakers are committed to providing liquidity to ensure that credit remains broadly available. As a result, the risk of this shock magnifying through financial market stress is limited and we don’t expect global financial conditions to tighten materially.
• The known unknown. It is important to recognize the limits of our knowledge about how the virus will spread. The underlying assumption is that the outbreak remains largely localized in China and fades through midyear. An alternative scenario in which it spreads widely outside China would materially alter the analysis presented here.
A large negative shock focused in China
With travel restrictions depressing spending in China and widespread factory shutdowns curtailing output, the COVID-19 outbreak represents significant negative demand and supply shocks. Over the past two weeks we have slashed our global GDP growth projection for 1Q20 in half to a 1.3% annualized rate. This would be the weakest outcome since the global financial crisis.
The biggest shock will be in China, where we forecast GDP growth to slow to a 1% ar this quarter from 6.3% pre-outbreak (Table 1). We expect the propagation of China’s weakness to be felt strongly in two groups. Countries in the region with close travel and supply chain links to China are likely to see exports weaken alongside a moderation in consumer demand (see here and here). In addition, commodity-producing countries should see growth weighed down by slower China demand and lower commodity prices. For the US and Western European economies we have taken a more modest 0.25% haircut to the current quarter, based largely on a forecast for slower export growth. DM demand is also likely to rotate this quarter as China’s factory shutdowns depress stock building while falling oil prices boost household purchasing power.
How deep is the ocean?
Although our forecast incorporates a sharp drop in China IP this month, it actually is based on optimistic assumptions. The baseline path for China IP assumes a 40% drop in production in early February, after which output quickly returns to full capacity by month-end (Figure 2). Thereafter, we assume production moves into overdrive and firms step up to restore February’s lost output entirely in the month of March. Under this baseline, China IP contracts 15% in February and jumps 28% in March (Table 2). The quarterly growth rates suggest a 4.2% annualized decline in 1Q20 followed by a 16.5% annualized jump in 2Q20 and an 8.3% annualized gain in 3Q20 before settling back to a more normal growth pace.
Behind this fast V-shaped scenario lies an expectation for a full “recovery” from China’s February production slump. A bounce back that merely returns production to its prior levels and pace of growth would fall well short of a complete recovery as the lost production would never be made up. This is unlikely if the shock is a temporary event. Thus, a true “recovery” requires production levels to overshoot prior levels for a period of time sufficient to make up for lost production.
The projected bounce-back recovery in our baseline looks too optimistic given the latest news (discussed below) that suggests the virus has done more damage than is in our baseline. We thus need to consider an alternative scenario that is a “delayed V-recovery” (Figure 2 and Table 2). Notably, a deeper downturn and longer recovery requires an even larger and longer overshoot to make up for lost production. The “delayed V-recovery” considers an initial 45% drop in output in the first week of February (5% more than the baseline) and a path of recovery that returns output to full capacity only by early April—a full month later than the baseline. As a result, the production overshoot, which begins in April, continues—and peaks—in May before slowing in June and only reaching a “normal” level of production in July. Under this scenario, IP plunges 34% in February, surges 34% in March, and jumps another 26% in May. The quarterly growth rates from this profile are significant, with IP collapsing 46% annualized in 1Q20 before soaring 227% annualized in 2Q and plunging again by 33% annualized in 3Q.
Tracking: Proof is in the data
The COVID-19 shock and projected impact on macroeconomic activity in 1H20 set us up for an extremely bumpy ride in terms of data tracking. The data over the coming days, weeks, and months will answer three key questions: 1) how deep is initial shock to China, 2) how long will the shock linger and how long will it take for a full recovery in China, and 3) how much will this shock spill over into the rest of world. In the optimistic scenario, the shock hits China production hard but results in a quick bounce-back with only limited spillover to the rest of the world. In the pessimistic scenario, activity in China falls even harder and takes well into 2H20 to recover, which in turn significantly disrupts supply chains and materially damps global growth this year. Our baseline is close to the optimistic scenario while the alternative scenario laid out above is more pessimistic on the China shock but maintains a V-shaped recovery that has a muted impact on global growth.
Which of these regimes China and the rest of the global economy is tracking will only be known with time and data tracking. Unfortunately, the monthly data are too lagged for tracking such a fast-moving story. Moreover, for China, the LNY will limit the availability of much of the key activity data to averages for January and February, which will not be released until early March. Given the magnitude of the shock and the importance of timing, we have turned up the microscope to track available daily and weekly data covering economic activity, air quality measures, travel statistics, tourism flows, and trade.
Currently, these data are sending a similar message of a sharp slowing in China beyond the usual moderation around the Lunar New Year (LNY). Daily coal consumption is currently running 40% below its average pace over the past decade (adjusted for LNY as well as average growth; Figure 3). Shipping volumes into China are also tracking a shortfall of about 30% relative to recent historical averages (Figure 4).
Perhaps even more important than tracking daily activity indicators, reports suggest factories are re-opening only slowly. Last week, the official pronouncement was for a one-week delay beyond the LNY to February 10. Progress has been much slower. Factories in the most affected regions remain shuttered, as the Hubei province just announced this week official shutdowns extending all the way out to February 21. Anecdotal information gleaned from company press releases by our European equity analysts confirms this more delayed return to production (Table 3).
Tracking: Spillovers in the monthly data
While the daily data are at the front lines of the tracking battle for China, the monthly data will be most important for spillovers to the rest of the world. February releases likely will bear the brunt of the virus shock with a sharp expected decline in global industrial output.
Our baseline forecast is that China’s manufacturing output will fall by 15%m/m in February, even with factories still to open next week. Assuming a modest 0.2% gain in January, this would imply a 7.4%m/m January-February average contraction in China—reported on March 16. Importantly, we do not expect such volatile swings in GDP. In the baseline, overall growth is projected to slow sharply to a 1%ar in 1Q20 and rebound quickly by 9.3%ar in 2Q20 before settling back to a 6%ar in 2H20. If the alternative path laid out above is right, GDP also would be somewhat more volatile but swings would still be quite muted given that we assume service sector activity will not be impacted as much as goods.
Outside China, we will gauge the spillover of the COVID-19 outbreak through our usual monthly indicators. If our baseline in right, global ex.-China factory output will contract 0.5% this month and rebound modestly by 0.3% in in March (Table 4). As a result, we expect global IP to contract 2.2% in February, its weakest outturn in nearly half a century outside the global financial crisis.
The first indications of the degree of spillover will come from the G-4 flash PMIs (out on February 21). We expect the manufacturing output PMI to drop nearly a full point in February. To the extent that most of our revisions to the global economy are in EM Asia and commodity-producing nations, the projection is unlikely to call for much of a drop in the G-4 PMIs. If the G-4 flash readings reveal a material decline, it would be our first indication that the spillovers to the major developed economies are larger than anticipated. Along with the G-4, Australia also produces a flash report. Given the country’s reliance on China for commodity exports, this report (also out February 21) will be a useful guide to spillovers. Beyond the flash reports, the full set of manufacturing PMIs are out on March 2. We expect the China manufacturing output PMI (IHS/Markit) to plunge 15pts, similar to the path seen in Japan following the Tohoku earthquake. Outside China (rest of world), we expect the manufacturing output PMI to tumble 49.7. On balance we project the global manufacturing output PMI to drop to 47.4, unwinding all the gains made since mid-2019. However, if our baseline forecast is right, and China and the rest of the world experience a quick V-shaped recovery, the PMIs should recover in March.
Global Economic Research