The shape of the rebound

Key points

  • The deterioration in the US labour market is steep
  • It seems the US is far from “peak epidemic”, consistent with a longer lockdown than in Europe, delaying the rebound in the world economy to the end of Q2.
  • We look at the likely shape of such rebound. We think it will be quite slow.

The epicentre of the pandemic first moved from Asia to Europe and is shifting to the United States which has now the highest number of covid-19 cases. There, labour market data suggest the economic toll is already very significant, even if the rise in jobless claims is to some extent a reflection of the generosity of the fiscal support brought about by the government.
In the very short run, we are quite relaxed on the capacity to protect income across all advanced economies, but the duration of the lockdown is the crucial variable. The later lockdowns start the longer they may have to last, since beyond a “critical threshold” at which the healthcare system is swamped out, time is needed to “re-charge” treatment capacity. While in Europe, and in Italy particularly, the news-flow on this front is positive, in the US it seems we are still far from “peak epidemic”. While we may tentatively pencil in a significant relaxation in lockdown measures by the end of April/beginning of May in Europe, we may have to wait until the end of Q2 in the US and given the level of trade integration this means we would have to wait until then before the normalisation of the world economy starts in earnest.
Beyond Q2, we need to start thinking about the shape of the recovery – assuming no relapse in the pandemic forces more lockdowns in Q2 and Q4. Informed by the Chinese data, it seems that consumers remain prudent upon exiting lockdown. Beyond behavioural changes, the decline in the value of accumulated financial assets and generic uncertainty may shift saving rates higher. But our main concern is investment. Uncertainty will take its toll there as well, but more fundamentally we think the rise in corporate debt which is likely to be unavoidable to protect spending on fixed costs such as labour compensation– on top of public spending – could impair capital expenditure.
There are currently many “recovery shapes” being discussed. For our part we would go for a “swoosh rebound” (the shape of the Nike logo), with positive, but fairly low GDP growth from Q3 onward.

Looking through the spectacular US labour market data

We have had quite a few spectacular economic data releases in the US last week. The deterioration in the labour market is taking place at a faster pace than what most observers were expecting, but this may be more a reflection of the specific features of the US policy response to the crisis rather than adding anything to the quantification of the recession, which we knew was going to be very significant anyway.

The market was expecting the “payroll data” for March to reveal a net loss of 100K jobs only, since the survey covered the first half of the month, i.e. before most states went into lockdown. The actual decline was 7 times higher, and the unemployment rate jumped from 3.8% to 4.4%. But as we suggested two weeks ago, when looking at “real time data” such as restaurant bookings, a large share of the US population “self-confined” early without waiting for government instructions (restaurant activity was already down by c/40% year-on-year in the week to March 15). It is thus no surprise that businesses hit by such spontaneous social distanciation started offloading workers quickly. The details of the release of the “establishment survey” confirm this: two third of the job losses came from the leisure and hospitality sectors.

Those payroll data are already completely obsolete compared with the weekly “jobless claims numbers”. Initial claims rose 6.6mn in the week to March 28th, after 3.3mn in the week before. There is always some discrepancy between the data used to build the unemployment figure (based on the monthly “household survey”, more on this below) and those jobless claims, but at first glance they are already consistent with a rate of close to 10% while the intensity of the lockdown measures has risen over the last few days, which suggests labour shedding is only beginning.  

Still, contrary to Europe where a number of “in-work unemployment benefit “systems exist and have been recently beefed up (“chomage partiel” in France, “Kurtzarbeit” in Germany), in the US there is not much of an alternative for employees to protect their income than applying for “full” unemployment benefits. In fact, the stimulus package adopted last week allows furloughed workers, and not just laid-off employees, to be eligible to unemployment benefits. The “furloughed” employees remain legally attached to their employers – the contractual relationship is not severed. Since the official international definition of being unemployed means being immediately available to take a job, the statistics ahead may not count those furloughed employees towards the unemployment rate.

Why is this important beyond the statistical nitty-gritty? Because looking ahead this may affect the shape of the recovery. Indeed, the depth and duration of mass under-employment after “peak lockdown” will partly depend on the strength of “attachment” between workers and businesses. Businesses will have to make a conscious decision to re-hire employees they have laid off, whereas resumption of “normal employment” should be swifter for those who opted for “furlough”. Also, furloughed employees usually keep the healthcare benefits provided by their employers, which – on top of the diminished uncertainty coming with retaining a higher probability to resume work after the crisis - should have a positive impact on their willingness to spend (no need to engage in precautionary saving towards potential health issues). In sum, the proportion of these two forms of labour shedding behind the rise in jobless claims in the weeks and months ahead will be one element to keep in mind when assessing the chances of a quick rebound in activity in the US.

What matters in the very short run is the shock to aggregate income, and on this front, we think we can afford to be relaxed. Federal support is quite generous, topping up the state benefits by up to USD600 per week for up to four months. On average, this means eligible employees relying on unemployment benefits will get the equivalent of the median in-work earnings (marginally south of USD 4,000 per month).

For now, household spending capacity is thus essentially unchanged. The government has substituted itself to the business sector to pay wages in the worst hit industries. We can work through the fiscal cost with simple assumptions. Federal Reserve of Saint Louis President James Bullard stated the unemployment rate could reach 30% temporarily in the US.  This would cost the federal government USD 120bn for a full month of indemnification, i.e. 0.6% of GDP.

The issue though is how long we will have to wait before the situation starts normalising and unfortunately, we are concerned with the “sunk cost” which the initial hesitant sanitary response in the US has triggered.

The later lockdowns start they longer they have to last

As of April 5th, the infection rate (number of cases per head) remained significantly lower in the US than in Italy (0.98 per 1,000 against 2.13 per 1,000) but as can be seen in Exhibit 1 the US growth rate – although slowing down – is equivalent to that of Italy as of three weeks ago. At the current daily growth rate, the US will have reached the current Italian infection rate in just 8 days. Another indication that the US is still quite far away from “peak epidemic” is that the ratio between “new cases” and “new tests” continues to rise there (see Exhibit 2), while it has started to fall in Italy. This would suggest that the healthcare system is still working through its backlog of cases in the US. Controlling for the difference in population, the testing effort so far in the US stands at 45% of the Italian level.

The first rationale behind lockdowns is to curb the flow of new cases to a level consistent with the treatment capacity of the healthcare system. From this point of view the US has no more room for manoeuvre than Italy when looking at the OECD data on “acute care hospital beds”, which stand at 2.4 per 1,000 in the United States against 2.6 in Italy (3.1 in France and 6.0 in Germany). Originally, it could be hoped that thanks to its younger population the US would face less pressure on its hospital system than in Europe. We should not overstate the impact though. Using the Worldometers data, the proportion of critical cases among all registered infections currently stands at 2.6% in the US against 3.1% in Italy. This is not a big enough difference to change much the level of stress on healthcare capacity, and the scramble to find additional respirators in New York City is very reminiscent of the scenes observed in Northern Italy a few weeks ago.

In Italy, fortunately pressure on hospital capacity is stabilising (see Exhibit 3) but this came a good three weeks into the national lockdown. The arithmetics are simple: once the epidemic has reached a “critical level” at which is threatens to swamp out the treatment capacity, even a drastic slowdown in the number of new cases leaves hospitals without any room for manoeuvre. This has obviously a bearing on the decision to relax the lockdown measures. Indeed, we need to see more than a mere stabilisation of the severe cases. We need to see a significant decline which would allow the healthcare system to “re-charge” in case another wave of infection were to emerge.

According to a report by “La Repubblica” on Sunday the Italian government has scheduled a discussion on relaxing the lockdown on some economic activity for April 14th, i.e. five weeks after the introduction of the national lockdown. And this relaxation would be quite gradual (starting with encouraging mechanical engineering businesses to re-start at full capacity, in order to make sure the main transport /logistical systems remain operational). Assuming no major relapse in the epidemic and based on the current “lags” across countries, we could tentatively pencil in a significant relaxation in the lockdown at some point in May in Europe as a whole, but we may well have to wait until June in the US, and as we have argued several times in Macrocast, given the level of trade integration the word economy will not start to normalise as long as all regions have not normalised.

Such timeline is of course highly speculative because there are still some major question marks surrounding the conditions for a normalisation. A key one is the testing capacity. This will to a large extent determine how quickly key workers can resume their activity, and while effort is being ramped up, it is still unclear whether all main Western countries (with probably the exception of Germany) will be in position to achieve this by the middle of Q2.

A swoosh-shaped rebound

Last week we discussed the magnitude of the shock to advanced economies during the peak of containment measures, and our baseline remains that a month-long lockdown would reduce quarterly GDP by 10%, and annual GDP by 2.5%. Given the recent Italian developments it’s probably safer to work with an assumption of a 5 to 6 weeks lockdown, but the key issue beyond this is probably how fast a rebound we should expect once logistical impairment is over.

Let’s start with consumption. We suggested in the first section that the impact of the lockdown on current household income would probably be minimal given the size of the fiscal support (this is particularly true in the US, but we think it applies in Europe as well). Since opportunities to spend are curtailed to some extent during the lockdown, this may create a temporary “hump” in savings. It is not obvious though that all this “forced saving” will be immediately spent when the containment measures are relaxed. The experience from China – where we have high frequency data - is quite telling from this point of view. Consumption on some key items is still significantly below normal: restaurant bookings are still a fraction of what they were before the epidemic and car sales are down 40%.  Of course, we only have this single observation, but at the very least we should probably assume that the “behavioural changes” do not disappear in the space of a few days or a few weeks.

Possibly more fundamentally we cannot exclude that the episode may have an impact on the preference for saving for at least a few quarters – or even permanently. Even if governments have been quick to come to the rescue, many households may have drawn the conclusion from the last few weeks that being able to draw on a “liquidity buffer” in case of massive disruption is a good thing. They may want to “stash away” the forced savings generated during the lockdown. We also need to consider the negative impact of the crisis on asset prices. In the countries – such as the US – where consumption is often quite sensitive to those, in any case some slowdown in spending will probably be seen, but the current demographic situation there may make this larger than usual. Indeed, we noted in the recent years in the US a rise in saving which could not be explained by the usual factors and we were tempted to ascribe this to the growing numbers of baby boomers coming close to retirement age topping up their pension funds. The destruction of value of the last few weeks may force an additional saving effort of this cohort.

But our main concern is with investment. First, mechanically the steep decline in capacity utilisation brought about by the lockdown - and after that some weakness in consumption - will have a negative impact on capital expenditure. Second, we think uncertainty will trigger a lot of wait-and-see attitude among global corporate decision-makers. Obviously, they will probably want to see in the second half of the year how the pandemic plays out, and crucially if we must go through more phases of lockdown. But they may also want to pause and re-think their industrial strategies, and in particular whether the “exposed fragility” of the supply chains would not warrant their shortening. The impact on investment locations could well be ambiguous. Would Western corporations choose to bring their production sites closer to their decision centers, or closer to their markets with the highest growth potential (which would mean not necessarily in Europe)? While this re-think is taking place, investment decisions may be put on ice.

We also suspect that the shock to the financial position of the corporate sector may leave lasting scars on investment.

Exhibit 4 suggests that investment was the great victim of the Great Recession of 2008-2009, falling by 20% year-on-year at trough, four times as much as the contraction in output.  Bear in mind that in 2008-2009 there was no sudden collapse in activity. In the US (but similar figures would be found for Europe) it took a full year for corporate output to reach its trough at -5% year-on-year, while we are likely to get there in just one quarter this time. But more importantly, at the time the root cause of the recession was an exogenous contraction in the supply of bank lending, as credit institutions were dealing with the deterioration of their own balance sheet triggered by the sub-prime crisis. As availability of external funding was drying out, corporations drew on their cash reserves. They burned 40% of their cash assets at the worst of the crisis (see Exhibit 5) as they were forcibly deleveraging. When the economy started improving towards the end of 2009, they immediately re-built their cash holdings – in an attempt to make themselves less dependent on external funding - while still reducing their debt level, thus constraining the rebound in capex and hiring.

In a way what the world economy needs right now is exactly the opposite. In 2008-2009 forced deleveraging triggered a contraction in spending on fixed costs – with labour cost in the front row – while what is necessary now is to make full use of the restored financial position of the banking sector to allow lending to protect current corporate spending, underpinned by government’s incentives (such as loan guarantees).  It’s in the collective interest to see debt-to-output levels rise.

This approach has a cost though. Indeed, in principle lending is used to raise productive capital. Of course some firms will miscalculate, but on average higher debt will be matched by higher profits, with financial sustainability intact. Today, we are asking businesses to raise funding to merely maintain current expenditure. This is consistent with a permanent deterioration in the firms’ financial position and thus their willingness to invest in the future.

The precise design of government support to lending in the current configuration is key. For instance in France the state guarantees apply to bank loans with reimbursement deferred by 12 months. This means that corporations will start paying back the loans raised during the emergency only when the economy has hopefully normalised, which is definitely a strong positive. Still, this is deferment, not cancellation, while the loss in activity at the peak of the crisis is very likely to be permanent. True the interest rate is going to be very small (the French government prices its guarantee to only 25 basis points on top of the banks’ own cost of resources, which is the basis for the interest rate on the emergency loans), but capital rembursements will still weigh on cash flows once the economy has exited the pandemic. The French government is ready to guarantee up to EUR 300bn, which would be enough to fully cover lost turnover for one month of the whole non-financial private sector (and we have some indication from INSEE that “only” a third of turnover is currently lost). But after the 12 month grace period businesses will need to pay this debt back in 5 years at most.  Assuming the entirety of the EUR300bn capacity is used, this means reimbursements would amount to c.15% of annual corporate margins for 5 years.

At least in France the system is already live and actively supporting businesses. In the US JP Morgan Chase stated last week that they would be unable to originate loans under a similar scheme promoted by the Treasury and the Federal Reserve because they did not have enough details on it. In the UK another state-sponsored system is in need of urgent re-think as uncertainty over the collateral that could be requested from borrowers blocked it.

This focus on corporate financial position calls for two types of action in the near future. One, it may be necessary, once our economies exit lockdown, to design specific fiscal measures to spur investment – specifically we think a temporary acceleration in amortization would help. Second, we think it would make sense to lengthen the duration of the emergency loans to make the weight on cash flow as small as possible in the short run (possibly with help from the central  banks).

In sum, with consumer spending possibly impaired by a higher propensity to save and serious potential curbs on investment, we think the likeliest scenario is for the world economy’s rebound post lockdown to be quite soft. World GDP would thus follow a swoosh shape (the Nike logo).