US Watch – Brace for the summer dip

by: Bill Diviney

  • Activity has recovered more quickly than expected following the easing of lockdowns
  • However, the resurgent pandemic is now leading to a stalling in that recovery
  • The net effect of this is to keep our base case intact, and we continue to expect the economy to contract by -4.8% in 2020, followed by a 3.4% recovery in 2021
  • While we see tentative signs of a peak in the second wave, the damage is already done
  • The policy reaction to the second wave has been much milder than the first; however, consumer caution is now a more significant driver of the renewed weakness in the economy
  • We are now also starting to see more powerful second-round effects, with permanent layoffs surging
200722-US-Watch-Brace-for-the-summer-dip.pdf (278 KB)

A rapid recovery that is now rapidly stalling

The economy has recovered more quickly than expected following the easing of lockdowns in May and June. However, this has proven to be a double-edged sword, and the resurgent pandemic is now leading to a premature stalling in that recovery. As a result, while the GDP report for Q2 released next Thursday will likely show a less severe contraction than we had initially predicted, the renewed stalling in the recovery has led us to revise down our growth expectations for the second half of the year; we now expect growth of 12.6% annualised in H2, compared with 18.1% previously. The net effect of this is that we maintain our base case growth forecast from early April, and continue to expect a -4.8% contraction in 2020, followed by a 3.4% recovery in 2021. Beyond that, as before, we expect continued consumer caution and second round effects of higher unemployment to continue weighing on the recovery in 2021, with activity still likely to be c.5pp below trend by the end of 2021.

Retail sales recover to near pre-pandemic levels

Consumption staged a surprisingly swift rebound in May and June. Retail sales surged in June by 7.5% mom, following an upwardly revised 18.2% gain in May. Remarkably, these cumulative gains took retail sales back very close to their pre-pandemic levels – total sales in June were just 0.6% lower than February, compared to the 21.7% shortfall in April. This occurred despite many states remaining partially closed, and indeed the eating out component of retail sales is still some 27% below pre-pandemic levels. Clothing sales also continue to be depressed. However, other categories of retail sales have – against our expectations – made up for these losses, with home improvement and recreation goods joining grocery sales in registering double digit growth. Online retail sales – for which we do not have a category breakdown – have also naturally surged during the pandemic.

Note that retail sales make up only around half of total private consumption (itself 70% of US GDP), which also includes expenditure on things such as housing, healthcare, education. These categories are the more ‘sticky’ components of private consumption, i.e. they were largely unaffected by lockdowns, and are less sensitive to overall macro conditions. However, there are some components of private consumption not captured in retail sales that have been heavily impacted by the pandemic, such as airline tickets, public transport, hotel stays, and other types of recreational services such as museums, cinemas and theatres. Naturally, the recovery in these categories lagged the recovery in retail sales in May, and we expect that to have continued to be the case in June. Even with these caveats in mind, though, total private consumption has still recovered more quickly than we had forecast – it was 11% below pre-pandemic levels in May (April: 18%) and is likely to have been 5-7% below pre-pandemic levels in June (this data is released on 31 July).

Industrial production and housing recover more modestly

Other areas of the economy have recovered largely in line with expectations so far, with the easing of supply side disruptions driving a pickup in industrial production to 11% below pre-pandemic levels in June (from a trough of -17% in April). Housing has also recovered; although housing starts in June were still quite subdued at 24% below pre-pandemic levels, the more forward-looking building permits have recovered more significantly, to 13% below pre-pandemic levels. We expect these recoveries to continue, as any renewed restrictions are likely to be more targeted at the riskiest hospitality sector rather than economy-wide, as in the initial lockdowns. However, the recoveries in these sectors will be dented by second-round effects of higher unemployment and weak consumer confidence, which will continue to weigh on demand.

Some of the consumption rebound looks unsustainable…

With the recovery so far being much swifter than we had expected, why are we not now raising our 2020 growth forecast for the US? Firstly, we think that some of the categories that saw surges in retail sales in May and June do not look to be sustainable; above-normal investments in home improvement, for instance, can only persist for so long. Meanwhile, this type of discretionary consumption had probably been spurred by the government’s $1200 stimulus checks; indications so far are that another round of transfers – if they happen – will be much more narrowly targeted at lower income households.

…but the impact of the resurgent pandemic is a much bigger cause for concern

However, a more significant reason for us to maintain our forecast is the resurgent pandemic, on the back of which there is now early evidence of a renewed fall in consumption and consumer confidence. While the hotspot states in the south look to have avoided the apocalyptic scenario of a return to full lockdowns – with the latest trends now suggesting we are approaching the peak of the second wave – the damage looks to have already been done, with renewed falls in indoor dining, a stalling in air passenger growth, and delays to further reopening – even in states less affected by the second wave (such as New York). These impacts are evident in higher frequency data. For instance, the OpenTable seated dining data recovered to as high as around 40% of 2019 levels at the beginning of July, before falling slightly and essentially flatlining since. TSA airport checkpoint data similarly shows air passenger numbers peaked in early July, and are now stuck at just 25-30% of 2019 levels.

This stalling in activity is being driven not just by renewed restrictions in the pandemic ‘hotspot’ states, but by changes in consumer behaviour. Indeed, the actions of state governors in response to the pandemic have been relatively modest – for instance, in Texas, the capacity limit for indoor dining was reduced from 75% to 50%, while in California, indoor dining was banned but outdoor dining continued to be allowed. This is despite the pandemic being much worse in these states in the second wave than it was in the first wave that triggered lockdowns – suggesting a markedly different policy reaction function compared to the early stages of the pandemic.

Policy reaction function has changed, but consumer behaviour a bigger driver

This is probably motivated by two key factors. First, the primary driver of lockdowns in the early phase – fear of hospital capacity being overwhelmed (which would affect not just covid-19 patients but anybody needing healthcare) – had been dealt with by boosting surge capacity in advance of the reopening, and indeed the new hotspots have managed the jump in hospital admissions much better than the states hit hardest in the first wave. Second, the state governors have likely observed that control of the pandemic can be achieved through a lighter-touch approach to lockdowns, with restrictions only on the riskiest of activity – as evidenced by the experience of some European countries (see Covid-19 hit to GDP: Some will fare much worse than others). They also possibly calculated that consumers would act more cautiously of their own accord in any case, following the previous experience with lockdowns. Indeed, research into mobile phone location data suggests that not everybody stayed at home because the government told them to do so, but because they chose to stay at home – perhaps through fear of catching the virus (See and

Meanwhile, it remains to be seen if we are really past the worst of the second wave, or if there might even be a third wave to come in the autumn/winter. Even if the pandemic is indeed peaking, it is a long road back to low infection rates in the hotspot states – perhaps 2 months – and this factor is likely to continue weighing on activity in these states, and potentially beyond.

Elevated unemployment and weak consumer confidence will also weigh on the recovery

All of this is without considering the second round effects from the first round of lockdowns, which are only recently becoming visible. While large-scale bankruptcies have at least been delayed by the myriad government support measures, permanent job losses are now surging – even if they are currently being disguised in headlines by the much bigger return of furloughed workers. Permanent job losses rose by 759k in June, up from 385k in May – the biggest loss in a single month since the height of the Great Recession in January 2009, when permanent layoffs were 805k. This took the stock of permanent layoffs to 3.7mn – the highest since July 2014, and so essentially erasing 6 years of labour market gains in the space of a few months. Unfortunately, we expect this trend to persist in the coming months, even though the broader unemployment rate is likely to continue to fall from elevated levels (11.1% as of June). We continue to expect unemployment to end the year in the high single digits – our forecast is 9.6%, with wide uncertainty bands. Rising permanent unemployment is likely one of the drivers of the recent fall back in consumer confidence, another of the second round effects that we expect to continue weighing on growth.