EsadeEcPol | Policy Insight
Authors: Jorge Galindo (Director of economic policy and data visualisation, EsadeEcPol), Antonio García Pascual (Visiting Scholar, Johns Hopkins University) & Toni Roldán (Director, EsadeEcPol)
1. The medical basis of the virus in Spain. Why do we think there will be more outbreaks?
The SARS-CoV-2 virus and its derivative disease Covid-19 have three characteristic traits that make this virus particularly harmful and difficult to control:
- Contagiousness. This virus has an R0 (average number of people infected by one person with the disease when no action is taken regarding the virus) of between 2.2 and 3.4 . The main aim of the policies to combat the epidemic is to reduce R to less than 1 so that each new infected person does not infect more than one other person.
- Mortality. According to recent studies and bearing in mind the considerable number of undetected cases, the virus apparently has the potential to kill between 0.2% and 1% of all infected persons.
- Severity. As a rough calculation, at the peak in Iceland (one of the countries carrying out the most complete tests), one in every hundred persons with the disease was in intensive care and around two/three in every hundred was receiving general hospital care.
On the basis of these data, basic epidemiological models anticipate that at least 60% of the population must be infected (or vaccinated) in order to achieve a reasonably effective level of immunity. Virtually no country in the world (including Spain) knows exactly how many infected persons it has for two reasons: many virus carriers have no symptoms – we do not know how many but some studies suggest 50% – and there are not enough test kits for the entire population.
Spain may have between ten and thirty times the detected number of cases
We therefore reckon that Spain may have between ten and thirty times the detected number of cases, i.e. a prevalence of some 6%-18% of the entire population.
This means that after the current flattening of the curve of infection and deaths, most of the population would still be vulnerable to infection. In other words, the current peak of the epidemic and the respective flattening of the curve only reflect the first outbreak of the epidemic.
This peak may be followed by others, forming a sort of epidemiological mountain range until a vaccine is created (or 60% herd immunity is reached).
The economic scenarios that await Spain will depend on how severe each new outbreak is and this will depend on Spain’s ability to mitigate the incidence of the virus.
Part I. Scenarios
In this section, we will examine three healthcare scenarios that entail significant changes for business and have different impacts on the economy.
On the basis of available data, we assume three external factors in all scenarios:
- No vaccine will be available or accessible until 2021 at the earliest.
- The summer may slow down aggregate contagion but will not prevent sudden outbreaks.
- The increased number of hospital beds and the development of antiviral medicine will reduce the severity of the virus but will not eliminate the risk of hospitals being overwhelmed by substantial new peaks.
Having defined the external constants (so far), we believe that the epidemic in Spain may span a spectrum of variable severity that can be summed up in three healthcare scenarios. The likelihood of Spain entering one scenario rather than another will vary according to how the virus evolves, and this will be affected by the decisions taken regarding operational and healthcare policies.
The government has begun to implement a lockdown easing plan in phases based on quantitative indicators that show how the epidemic is evolving, in keeping with the recommendations made by EsadeEcPol. It is, however, essential to emphasise that the sustained success of lockdown easing during the provisional schedule put forward (according to which Spain would resume everyday business in a manner resembling controlled normality around July 2020) depends entirely on the ability to develop ways of measuring the epidemic, tracking contacts and practicing self-isolation. Until such capabilities are assured, we feel that the outcome of the plan is still uncertain.
Each of the three scenarios is assigned an impact determined by two correlated dimensions: mobility and economic activity.
- Mobility of the population is the main dependent variable affected by an epidemic of these proportions (in addition to the healthcare dimension itself) not only due to the decisions taken to contain it and thus reduce the infection rate, but also due to the decisions taken by private agents in response to risk and uncertainty. Aggregate data from mobile platforms (Apple, Google) in recent weeks in several countries around the globe enable us to estimate the expected reduction in mobility in each instance.
Both platforms provide these data as a percentage reduction in activity compared to the average recorded in January. Graphs of data are produced daily, divided into categories. Those used by Apple refer to the sort of transport (on foot, private vehicles, public transport) whilst those used by Google focus on the purpose of each trip (leisure, essential shopping, commuting and travel being the main ones). We will focus on the use of private vehicles (henceforth Apple) and commuting (henceforth Google), because they give the best insight into the impact on the economy.
Extrapolation, by comparing moments and countries in situations similar to each scenario, makes it possible to estimate their impact on mobility. All these data were accessed on the 25 April 2020 baseline, therefore that date is the cut-off point of the data given unless stated otherwise.
- Economic activity. We have created three macroeconomic scenarios based on the healthcare scenarios and their expected impact on mobility, and adjusted them using data from the abovementioned platforms. We took a pre-Covid scenario as the benchmark that defines the usual monthly level of activity, and calculated the percentage of activity in each sector as a function of the impact of self isolation and social distancing on each sector. These were then aggregated until an activity percentage was obtained for the Spanish economy that evolves month by month according to the healthcare scenario and its impact on mobility.
The OECD, for example, estimates that in the latter half of March, the average activity in OECD countries was 70%, i.e. lockdown and social distancing caused a reduction of 30% in comparison with normal activity in March. We estimate that activity in the second fortnight of March in Spain was close to 60%, i.e. somewhat worse than the OECD average for reasons that we will explain shortly. Bearing in mind that activity in the first two weeks was close to 100%, we calculate the March aggregate to be 80%. Thus March reduces 2020 growth by almost 1.7%. April had more weeks of lockdown and the level of activity fell to 70%. This reduces growth by another 2.5%. As the lockdown in our healthcare scenarios is gradually lifted, the level of activity improves to 90% in December.
Figure 1. Economic growth in Spain. Forecasts by EsadeEcPol.
Healthcare assumptions: new, semi-contained outbreaks in 2020
The plan for lockdown easing in phases announced by the government does not prevent the recurrence of partial new outbreaks at specific times and areas of Spain that lead to regionalised, intermittent partial lockdowns in 2020.
In this scenario, we assume that in the absence of sufficient testing, measurement, tracking and self isolating Spain will be unable to keep R under 1 at all times across the nation and must therefore implement an asymmetric strategy to pin down the outbreak as it occurs: greater epidemiological capacity (but not enough) is combined with unavoidable, strict and rather indiscriminate social distancing. Furthermore, social distancing will remain necessary even after lockdown.
The plan for lockdown easing in phases does not prevent the recurrence of partial new outbreaks
One case that would help us estimate today the mobility outcome of the future is Singapore: after containing infection for more than a month, Singapore was obliged to impose a general lockdown at the beginning of April. Alternating lockdown and no-lockdown periods caused an equally high variance in the population mobility estimates produced by Apple and Google: from a 20% reduction before lockdown in commuting according to Google and around 30% in private vehicle aggregates according to Apple, to levels of 60%-65% in both.
With intermittent lockdowns, mobility is expected to lie between the two values. Assuming that the country is not in total lockdown most of the time (at least two thirds), then a reasonable estimate is a reduction of around 40% in aggregate mobility. Other current instances of non-mandatory or fragmented semi-lockdown (Florida, USA; Mexico City; Sweden) have values of around 35%-45%, which bears out the hypothesis.
Economic impact: 15% reduction in GDP in 2020
Social distancing and new outbreaks in the latter half of the year have an enormous impact on the economy in 2020. In addition, the return to normal life does not happen until 2022. We modelled the economic impact sector by sector and week by week, building upon the basis of the scenarios, week by week, of mobility affected by lockdown and social distancing.
Social distancing and new outbreaks have an enormous impact on the economy in 2020
In this way, the greatest impact on growth occurs in the last weeks of March and the first weeks of April. Aggregating bottom up from the weekly to the monthly impact, and from the impact on individual sectors to the economy as a whole, we reach an average level of activity in the first few weeks of lockdown (the latter half of March and early April) which in the case of Spain is 60% of the normal level (i.e. with no pandemic). This estimate is 10 percentage points lower than the average level of activity in developed countries (OECD).
Working on the basis of the healthcare scenario, we assume that the gradual improvement in the level of activity is disrupted by a new outbreak at the end of the third quarter (less severe than the March/April outbreak). Activity would shrink from 90% in September to 80% in October and bounce back to 85% in November and 90% in December.
The aggregate impact would result in a business level of 85%, i.e. a reduction of almost 15% in GDP in 2020 (Figure 1) and an upturn of 12% in 2021.
Figure 2. Annual decline in Spain's GDP.
After the aggregate impact on GDP, there are several considerable differences between sectors that vary according to the "contact level" of each sector. Figure 2 shows how the decline in the service sector in 2020 (-16%) is twice that of industry (-8%). Meanwhile, the “Retail, Transport and Hospitality” area of the service sector would shrink by 31%, almost twice as much as the sector average due to being "high contact" business.
The crisis is hitting Spain harder than the OECD average mainly for two reasons.
- Firstly, the health crisis is more virulent: the numbers of infections and deaths have situated Spain, together with Italy, at the European epicentre of the pandemic.
- Secondly, the structure of Spain’s economy is more vulnerable to Covid-19 than other EU countries due to its high numbers of SMEs and self-employed, the importance of the tourist industry and "high-contact" sectors, and also the excessive numbers of temporary employment.
As a matter of fact, the (preliminary) economic data we already have for the first quarter enable us to compare the Spanish economy in the early weeks of the pandemic with the Eurozone average. The GDP figures for the first quarter suggest a drop of 5.2% in comparison with the previous quarter in Spain, the worst figure in recent times, whilst the eurozone shrank by 3.8%.
The impact will, of course, be far greater in the second quarter because the lockdown affected the economy from the beginning of the month and not just the third week onwards as occurred in March.
We estimate that unemployment during lockdown will peak at above 22%
We also already have some employment figures which reveal the slump in job market conditions: a sudden halt in the number of hours worked since the lockdown began. This is reflected in the rise in unemployment and the fall in numbers in employment, although thanks to state subsidies and temporary collective redundancy schemes that maintain employer-employee relations during the lockdown, the impact has been cushioned to a certain extent.
We estimate that unemployment during lockdown will peak at above 22% (at the beginning of the second quarter). If the temporary collective redundancy schemes are included, unemployment will soar beyond 30% at the lockdown peak, the highest level since records began.
Best case scenario
Healthcare assumptions: epidemic under control. The lockdown easing put forward by the government can be carried out and maintained due to the relative control of the epidemic with R<1 thanks to the increased capacity of testing, tracking and self isolation in addition to the implementation of more widespread and sustainable physical distancing plans, with a high impact on R, at an affordable mobility cost.
The lockdown easing can be maintained due to the relative control of the epidemic
In countries managing to contain the epidemic without resorting to lockdowns, decreases of around 20-25% in aggregate mobility persist according to Google and Apple. This was the case of Singapore before it had to resort to its own (latter half of March according to Google); and Taiwan according to Apple at its lowest points in the last-but-one week of April; and South Korea throughout March and April.
Impact on the economy: a drop of almost 9% in GDP. In this scenario, the impact on the economy is limited to the first half of 2020 and part of the third quarter thanks to an effective strategy in the plan for easing the lockdown and containing the pandemic.
GDP shrinks by 9% in 2020, and grows by 6% in 2021. In this scenario there is virtually no permanent damage to the economy. Although the volume of business in some sectors, such as tourism and other high-contact sectors, will remain below pre-Covid19 levels even in 2021, prompt, effective action will minimise any permanent damage.
Worst case scenario
Health-related assumptions: another outbreak of the March epidemic. Either the lockdown easing stages are not completed or they are completed and then have to be reinstated due to uncontrolled outbreaks causing at least one upsurge in the epidemic similar to the one already experienced (in terms of numbers of cases, deaths and the possibility of overwhelming the healthcare system), leading to another lockdown of the same proportions as the current one.
Travel in private vehicles and commuting plummeted by more than 75% during lockdown in Spain and Italy. It seems reasonable to expect reductions of the same order of magnitude due to the indiscriminate implementation of collective confinement to contain the virus until the end of 2020.
Impact on the economy: 20% drop in GDP. The macroeconomic impact would be much more severe with GDP shrinking by 20% in 2020 due to new, more virulent outbreaks in the latter half of the year that would lead to strict new lockdowns and cause the economy to fall to below 70% for six weeks between September and November (Table 1).
Table 1. Evolution of Spain’s GDP in 2020 by month according to volume of business possible during the epidemic
Persistently low levels of activity would also cause companies to go bankrupt and jobs to be lost, increasing the likelihood of certain banking entities having liquidity problems. This would aggravate and delay recovery even more.
The big question and risk in this scenario is the medium-term impact of the pandemic on productivity and growth. Lower growth and the prolongation of tax breaks for the private sector would increase borrowing and fiscal deficit more than other scenarios. Considerable EU support for member states would be essential for recovery.
Part II. Operational policies
In the event of an outbreak, the ideal epidemiological response would be to:
- Know exactly at all times who (a) is infected and could pass on the virus, (b) has been infected in the past and acquired temporary immunity and cannot pass the virus on, and (c) is not infected and is therefore vulnerable to infection.
- Know exactly who has been or is in contact with infected persons.
- Be able to confine group (a) individually, and enable groups (b, c) to engage in everyday interaction.
If the epidemic overwhelms the authorities’ capacity to comply with any of the three requirements, tougher measures will be necessary.
As we described in the decision flow chart in our article Medir, rastrear, aislar, the closer we are to complying with it, the greater our likelihood of entering a best case scenario. The further away we are, the more likely the worst case scenario will be.
Action necessary for Spain to approach the best case scenario
1. Capacity to measure, monitor and isolate individual cases
Acquisition of more high-standard tests and optimisation of their use, giving priority to periodic (in successive series) and regionally representative surveys to detect the incidence and spread of the virus by seroprevalence testing. The government is conducting a survey of this type to measure the incidence of the virus at different regional levels. This survey is not only essential in order to estimate the scope of the epidemic, it should in fact have been carried out before and, furthermore, it should be repeated periodically to incorporate different regional decisions.
RT-PCR tests do not distinguish between people who have been infected and are now immune and people who have never had the virus
As regards optimum testing capacity, we do not have enough data to evaluate the situation in Spain. At present, RT-PCR tests are the most widely used. They are accurate (particularly the traditional, not quick tests) but do not distinguish between people who have been infected and are now immune and people who have never had the virus.
The government has been increasing the testing capacity by means of seropositivity testing, which is appropriate because it distinguishes between the three groups mentioned earlier. However, the tendency of some tests to return false positive results calls for caution when choosing between the alternatives on the market. Furthermore, precisely for this reason, it would not be advisable to use them as a mechanism for issuing travel permits.
- Digital platforms based on the capacity provided recently by a partnership between Google and Apple (together they account for virtually all the mobile devices in Spain) to create bluetooth-based logs of infected persons. On the basis of an opt-in, decentralised model, such logs would enable anonymous isolation alerts and recommendations to be made when an uninfected person comes into contact with a contagious person. In this respect, European integration would be not only desirable but necessary in the context of the theoretically open-border Schengen Area, but so far progress has been limited.
- Massive increase in the workforce monitoring the epidemic in Spain, by taking up available capacity in public administrations and hiring staff if necessary (which would be an additional stimulus policy).
Design of effective self-isolating protocols to minimise contact within households when physically feasible, and making areas with provisions for self-isolating available nationwide.
2. Design of a precise lockdown easing model
The input of this model will be based on early alert indicators and its output will be the isolation measures. The government’s current lockdown easing plan is based on this model and puts down on paper the intention to deal with the uncertainty inherent in a pandemic of this nature.
Is is essential for the government to ensure the use of accessible, reliable indicators
It is, however, essential for the government to ensure the use of accessible, reliable indicators: (1) periodic seroprevalence surveys; (2) specific thresholds regarding ICU admissions and aggregate deaths, not only those identified as Covid-19 positive, compared with the average of previous years. This removes the need to rely on diagnostic tests, which are scarce in Spain and elsewhere in the world.
At the same time, it is essential for the mandatory isolation measures, triggered when indicator thresholds are exceeded, to be chosen according to the following criterion: the greatest reduction in the contagion rate at the lowest possible socioeconomic cost.
Insofar as suitably regionalised data are available, the measures could be implemented in different ways in limited areas (cities, provinces, autonomous regions) according to a zoning model like the one put forward by EsadeEcPol. The province-based zoning currently used by the government seems to be of suitable size, with the added advantage that it is an easily recognised benchmark for the population.
A modular isolation framework dependent on visible indicators also has the advantage of enabling private agents (individuals, households, families, companies and other entities) to adjust their expectations and get ready for the eventualities of each phase, reducing uncertainty as much as possible and, therefore, slightly lessening the economic impact of living with the epidemic.
Part II. Economic policy during the transition period
So far, the tough, early response of the ECB has managed to stave off, for the time being at least, the more vulnerable countries in the Eurozone taking a heavy hit from the markets. But the ECB should not be the only European institution in the battle against the pandemic. As a complement to the ECB, a package has been proposed consisting of immediate aid for the states (ESM), workers (SURE) and companies (BEI) in the form of loans and guarantees amounting to €540bn.
In addition, in order to boost recovery after the pandemic, member states are preparing a common European fund that might provide €1.5tr for aid and investment. These proposals are promising but need to be more detailed and better designed.
- The terms of ESM loans to combat the pandemic are not generous enough. In order for ESM loans to have sufficient clout and contribute to the sustainability of public borrowing, they should have a margin of zero or very close to zero, be long term (ideally longer than 15 years) and equivalent to about 10% of GDP. Current conditions (2% of GDP, a margin of more than 10 base points, and a term of 5 to 10 years) are not generous enough, particularly bearing in mind that the BCE, particularly thanks to its security buying programme, already reduces the cost of finance for member states.
- The financing and use of the European recovery fund are yet to be defined. How the European recovery fund is to be financed and the sort of aid or investment it will be used for are crucial for its success. Ideally it will be important to agree upon a considerable degree of mutualisation via shared debt or, better still, shared equity to finance European projects. The projects should focus on strategic sectors such as R&D, healthcare or the environment. The shared equity would provide the greatest degree of mutualisation which could benefit all states. In addition, the distribution and location of the investments could take into account the extent of the impact of the pandemic.
In mid March, the Spanish government passed an unprecedented package in response to the crisis, in line with other European governments. Since then seven royal decrees of measures focussing on three main cornerstones have been passed:
- Facilitate companies' access to loans by means of guarantees and public credit.
- Cover a good part of companies' wage bills (and other costs) to prevent them from going bankrupt.
- Provide protection for the most vulnerable workers such as the self-employed, temporary workers and domestic workers.
In general, the measures are moving in the right direction although in the future it would be advisable to take two key matters into account.
Ensure that the aid is timely
The government has aimed "not to leave anyone behind". However, when designing social and economic policies, little attention has been paid, unlike in other countries, to criteria of flexibility and simplicity regarding the benefits and support that are so essential in times of emergency. As a result, many workers and companies are not in fact getting the necessary support.
Benefits often take months to materialise
In normal times, usual benefits, logically enough, involve complicated bureaucratic procedures beforehand to ensure that the recipients are the ones who really need them. These drawn-out, laborious procedures are ineffective now, at a time when help is needed immediately and the administration is only providing minimum services due to the virus.
Benefits often take months to materialise. We must design simpler, more universal schemes with fewer requirements. Designing measures tailored to each vulnerable group, as Spain has tried to do, also runs the risk of sidelining many people who need benefits. At times like this it is better to go too far than not far enough and suffer a deeper recession.
Maximum flexibility and incentives for using temporary collective redundancy schemes for longer
The experience of other countries shows that, in times of crisis, systems such as the German Kurzarbeit or similar schemes allowing temporary changes to working conditions and some labour costs to be covered by the state, enable job losses to be reduced significantly.
Spain’s job market is a distinctly dual market – with a high rate of temporary employment and ineffective internal adjustment mechanisms in companies – that traditionally causes massive job losses during crises. In the current crisis, however, despite the unprecedented slump in the number of persons registered with the Social Security system in March and April, the widespread implementation of temporary collective redundancy schemes has enabled the impact on unemployment to be significantly reduced.
The severity of the economic fallout will depend directly on the healthcare scenarios
During the transition period, it will be essential for regulations to be modified to enable companies to apply them with the greatest possible flexibility, and be designed so as to avoid excessive disputes and facilitate administrative procedures because companies will continue to face difficulties related to the Covid-19 crisis for many months.
The severity of the economic fallout will depend directly on the healthcare scenarios. In our case, our starting point is a baseline scenario that is somewhat more pessimistic than the consensus in the latest estimates made by the government, the IMF and the Bank of Spain because our healthcare assumptions are somewhat more negative.
Spain does not have sufficient operating capacity to implement a comprehensive measurement, tracking and self-isolation strategy on a par with those of other countries more accustomed to dealing with respiratory epidemics. If countries like Singapore and South Korea which do have such mechanisms are finding it very difficult to deal with each new outbreak, the risks of each new outbreak in less well prepared countries such as Spain are enormous.
The government must focus all its efforts on building the capabilities needed to live in a "low-contact economy" and providing all the support necessary during the transition period in the most flexible manner possible.
- Lai, CC, Shih, TP, Ko, WC, Tang, HJ, & Hsueh, PR. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and coronavirus disease-2019 (Covid-19): the epidemic and the challenges, Int J Antimicrob Agents, 55(3):105924 (2020)
- MRC Centre for Global Infectious Disease Analysis. Report 9 - Impact of non-pharmaceutical interventions (NPIs) to reduce Covid-19 mortality and healthcare demand, Imperial College London (2020)
- Gideon Meyerowitz-Katz. What Is the infection-fatality rate of Covid-19? (2020)
- Carl Heneghan, Jon Brassey, Tom Jefferson. Covid-19: What proportion are asymptomatic? CEBM (6 April 2020)
- Kiko Llaneras (2020). Números del coronavirus: cuatro preguntas clave antes de recuperar la vida normal, El País (26 April 2020)
Join the Do Better community
Become a member and enjoy our free benefits. Get recommendations, receive personalised content in your inbox and save your favourite articles to read later.