C-19: Why most economic projections have likely got it wrong

Updated: Jun 9, 2020

An understanding of how the economy has performed during the period of lockdown has only just been made clear, with company quarterly reports and employment statistics now out and the market quickly digesting and 'correcting' prices. By and large a gradual correction back to pre-pandemic levels is evident across the board - the NASDAQ index (.IXIC) is in fact already back to same levels as it was in early February. There is further stretching of, and distancing from true reality, as PE ratios continue rising and stock market-breadth indices weaken; coinciding with a few very large [tech] stocks buoying major markets despite revenue and profitability for players diminish across the board due to a rapidly weakening economy. However, investors are not scientists and epidemiologist and, quite frankly, markets are absolutely rife right now with speculation as to how the crisis will evolve and how economy will pan out. In this article we attempt to link the reality from the pandemic on the ground to how economies may evolve and perform. In doing so, we develop a 'low resolution' and rudimentary model by stripping back to a few basic fundamentals, to determine how the virus will continue to affect economies and therefore their future performance. Read on to find out more.

Update 7 June: a follow up to this article is available here.

1. Introduction: what most are predicting as the economic outlook for the world economy

  • In general, one thing is clear across most projections: there is little hope and consensus for a V-shaped recovery. At best a U- or W- shape is what most are expecting and can harbour hope for. That is, a long hard slog battling the virus back and forth until a vaccine eventually emerges. The figure below shows the IMF's projections for GDP outlook for 2020 and 2021 across a range of countries and regions. Other reputable projections look relatively similar, e.g. the WorldBank's.

  • What typifies these projections seems to be an extrapolation of how countries have fared thus far, from the initial phase of dealing with the pandemic. To recap, this is the phase where lockdowns featured heavily as the primary intervention to avert and flatten the curve of exponentially rising deaths and infections. Most of us are quite familiar with how this went: a shocking speed of spread in Europe and North America, resulting in massive deaths in the order of ten's of thousands per country, in contrast to a relatively muted impact to Asian countries outside of China.

  • As we had pointed out in this earlier article here, the current second phase of the epidemic spread has a very different objective function compared to the first one: keeping transmission rate below epidemic levels (<1.0), while allowing for as much economic activity and physical movement among the population as possible. And to reiterate what we have said about this before: what happened during the first phase of lockdowns has little to no bearing on how countries will fare in the second one.

  • In other words the current economic projections may have it wrong, as they build on an understanding about the past, but an inadequately robust understanding about the possible future evolution of the virus in each country. This is what we intend to investigate and understand in greater depth, and hopefully correct for in this article.

Figure: World economic outlook projections from the IMF (April 2020). Visibly negative 2020 GDP projections across the board with exception of a few (e.g India and China), reaching some -7.5% for the Euro zone, -5.9% for the US and between -0.5 to +2% for emerging Asian economies.

Source: https://www.imf.org/external/pubs/ft/weo/2020/01/weodata/index.aspx

2. Could current market performance be a credible indicator of how the economy could evolve? Our answer: in short, likely not.

  • In theory the market should provide reliable feedback on the state of the economy. Either as a representation of how companies are performing in the current environment, or as a forecast of future performance given market conditions and outlook, or both. At present, it is unclear whether either is in fact the case.

  • There has already been much discussion on how overvalued the US stock market was coming in to the end of 2019, when price/earnings (PE) ratio for the S&P 500 hit a level of around ~19x. Today's situation of suppressed earnings, given the pandemic, has pushed the same metric to 23.5x versus a historic median since the 1900s of ~15x (an unfathomable premium of some 55%). One can repeat this computation with the Shiller method to adjust for cyclicality of the economy and still end up with a similar result. The bottom line is: markets are likely highly overvalued at this point in time, and made worse still by the on-going situation of the pandemic.

  • We are not here however to discuss market valuations. Simply, there is only one point we want to get across in this section: that market prices are a relatively poor indicator of how the economy is performing and where it is going.

  • Several additional points underpin this viewpoint. (1) There is no reason for markets indices to move almost exactly in tandem with one another, yet that is what we see consistently across markets (see figures below for select indices and their cross-correlations); (2) Inconsistency in index changes vs. COVID-19 reality: market index levels for some hard hit countries (e.g. the US, which can be partially represented by the NASDAQ index or .IXIC) have almost fully recovered to base levels, while less affected ones like Germany (represented by the DAX index), and South Korea (KOSPI) are still some 15-20% from base levels as at mid-May; (3) Large sways and movements in markets are often underpinned by the actions of several very large asset managers - i.e. the Blackrocks and Vanguards of the world. Given the market structure in the investment space, these players deploy capital at the clip of billions of dollars to worldwide marketplaces, and often buy and sell assets in individual markets merely as a means of adjusting indexing levels. In short, a few major players may be primarily dictating how prices and indices evolve, not real aggregated sentiment of a long tail of investors where perhaps the law of large numbers has a fighting chance to prevail.

  • For further insights, this video from CNBC discusses the gap between the stock market and broader US economy - link.

Figure below: Daily market indices for a number of markets around the world; including US (DJI, IXIC/Nasdaq, RUT/Russell), UK/Europe (RDX, UKX, EWL, AEX, IBEX, PX1/CAC, RUT), Asia (SENSEX, KOSPI, NIFTY, NI225/Nikkei, IDX) and China (SSE). There is clear strong correlation across different markets (in timing and magnitude of changes) despite widely differing situations on the ground from country to country.

Figures below: (a) Weekly prices for select ETF assets representing different asset classes; across geographies, equities, bonds, commodities, indexed to first week February 2020 - we covered the initial impact of pandemic on the prices of these assets in an earlier article here; (b) Cross-correlation between these assets over the same period. As expected, we can observe a highly positive correlation across equity assets (regardless of market), and low/negative correlation to certain asset classes such as commodities and bonds - consistent with the hypothesis that the market may not be adequately reflecting the true realities of markets that distinctly vary from one economy to the next. Instead, it is indicative of flights to safety by toggling between major asset classes and showing visible herding behaviour across similar asset classes.

3. We attempt to construct a synthetic model of how the economy may evolve from this point onwards by distilling into four key factors that may matter most

  • This is the crux of this article, and perhaps a slightly tall order to attempt. To start, an understanding of how the pandemic has affected and will continue to affect any one country should directionally provide a more robust truth compared to what the economic 'experts' and investors/prospectors have forecasted.

  • The first step is to break-down the individual components of potential driving factors that are likely to effect a sizeable deviation to the economy from its original base trajectory. This can be split into the effects from the first phase (1) the lockdown: a finite period where economic and trade activity has been practically obliterated, as populations are sheltered in place and a large proportion of business activity put on hold for some 2-3 months. Followed by the second phase; which we can further break down to three components: (2) travel & tourism related: effects from sequestered industries that have little to no hope of recovery due to long-standing physical distancing, (3) prolonged restriction: effects from continued inhibited and restricted movement of the population that continue beyond lockdown, and (4) trade impact: effects from reduced trade related economic activity that are dependent on other trading partners.

  • The effect from lockdowns (1) should be relatively consistent across all countries - governments and businesses alike have been largely focused on handling the initial effects brought about by the pandemic, and understandably, most of the working population was paralysed from going about their daily lives. Perhaps markets were not far off here in their assumptions, a sharp 25-40% drop to the lowest point was common across major market indices during mid to end March 2020 period. Presumably, this was in line with the actual economic impact arising from the lockdowns.

  • Update 7 June: US live spending data sourced from Exabel suggests slower than expected recovery from lockdown activities. While mobility recovers with easing of restrictions, spending is slower to recover back to baseline levels. As at 30th May, combined spending is still -30% from baseline levels due to lockdowns initiated around mid March 2020, with a lowest point of -40-45% at end March. In contrast, our composite mobility level index registered -50% at lowest point, but reached around -20-25% by end May.

Figure: live transaction data from the US showing spending activity across different sectors (period Jan 20th to May 23rd 2020). There is a clear and persistent decline in spending across most sectors (inconsistent with 'market' information), while some recovery is happening it is very slow and still impacting combined spending across sectors very significantly.

Source: https://1010data.exabel.com/covid-19/

Figure: mobility trend lines for the US across major mobility types. Broadly consistent with spending patterns indicating impacts of lockdown, however in spite of recovery in mobility, the upwards trending is not translating to spending as quickly

  • For travel & tourism related factors (2); across countries, these should affect industries such as tourism, travel, and airlines consistently and to a similar extent. The overall impact will however vary at aggregate level from country to country given different levels of dependence on these industries, as well as differences in dynamics for local indigenous tourism. Travel and tourism economic data in the figure below may help us better understand this - by and large the broad sector contributes 7-14% to total country GDP, a fairly significant proportion.

  • Prolonged restrictions (3), is where large differences could emerge from one country to the next, as for some, the crisis will be prolonged significantly - as we have discussed in a previous article here. To add, private consumption today is a large part of the GDP in almost every economy, ranging between 45% to 85% - see figure below for reference data. As mobility and physical interactions are inhibited to the point where economic activity suffers (as had happened during lockdowns), the downside impact on the economy will be tremendous. The longer a country needs to stay locked down and formally or informally inhibit physical interactions, the worse this is. As we have touched on before, there are countries that where this outlook is extremely worrying; namely Mexico, Bangladesh, Egypt, Brazil, India, Indonesia and Philippines.

  • Trade impact (4) is where the effects due to changes in imports and exports to and from partner countries will manifest. In effect these are the secondary and tertiary effects of the activities in factor (2) and (3) combined, i.e. where there is related to trade activity with partner countries.

  • The following two figures provide some background data that could be useful as reference.

Figure: 2019 GDP contributions on travel and tourism by economic region, as reported by the World Travel and Tourism Council

Source: https://wttc.org/Research/Economic-Impact

Figure: consumption, exports and net exports as % of GDP for a range of countries (2018 data). Source: Worldbank statistics database

4. Making an educated guess of the future economic trajectory for a given economy. Crucially, accounting for potential continued physical restrictions that may continue beyond lockdowns.

  • Sections 2 through 3 were perhaps a longwinded way to suggest how and why current economic forecasts are not fully robust, or at the very least misguided in their ability to comprehensively foretell the future economic outlook. In our humble view, none seem to credibly provide adequate clarity on how the actual progress made by each country in dealing with the virus thus far may then translate to impact on economic activity.

  • We can start by formulating the computation of the effects of each of the four factors described. To recap them here: (1) the lockdown, (2) travel & tourism, (3) prolonged restriction(s), and (4) trade impacts. Bear in mind, not all these factors will effect each country equivalently. Namely, travel & tourism, as well as trade impact depends on the dependence of a country's economy on those activities. The following paragraphs explores each factor in more detail, while a figure included below helps summarise them in one place.

  • (1) Across the world, private companies particularly in the consumer space have widely cited a drop in revenue ranging between 25-40% from lockdowns. UK retail sales dropped 5.1% in March, with clothing store sales falling as much as 35% [FT]; the aggregated negative trend buffered only by the fact that panic buying and stockpiling increased sales for food items and household essentials running up to April. One consultancy in the UK estimated each month of lockdowns could cut 1.5% off annual growth [link]. This obviously varies from country to country and by sector, but gives us a general sense of its magnitude. Assuming a 1.5-2% impact during lockdowns, with length ranging between 3-4 months, an annualised overall economic impact between 5-8% can be expected.

  • (2) With travel and tourism contributing some 7-15% of economic impact from country to country - a hypothetical estimate of 9-12 months of strict travel restrictions, lasting until standard operating procedures are put in place and fully implemented, could result in an annualised impact of 5-10% to overall GDP. Admittedly, this is a vastly simplified computation that excludes multiple factors such as local/indigenous tourism, and international travel correlation to business dealings and the like. However, the key factor here is perhaps the psychological impact of the numerate deaths from COVID-19 so far, that will not be so easy to be forgotten without the aid of time. It will take a significant amount of time for most people to feel comfortable travelling again.

  • (3) Prolonged mobility restrictions depend on whether the country is progressing well in its second phase efforts to maintain transmission below epidemic levels. We described this at length in our previous article on the topic. Several countries are expected to suffer protracted crises here, while many others have little to worry about but will nonetheless insist on taking baby steps. Hard hit countries in Europe like Spain, Italy, France, and the UK have fortunately, but also surprisingly, been able to buckle the correlation between mobility and transmission through non-movement restricting interventions [link]. Today, transmission rates already stand at vastly lower levels to where they stood before lockdowns, even at the same mobility levels. In fact Germany, Austria, Italy, Netherlands and Switzerland are less than a month away from 100% mobility, and not far behind are Belgium, Spain and France - crucially, without invoking epidemic transmission rates. Meanwhile, this is far from the case for populous countries like Brazil, India, Egypt, Indonesia, and Philippines. Transmission rates continue to struggle to stay at the 1.0 mark even at low mobility levels, yet mobility continues to creep up by the day. It is expected that this struggle could continue on for another 3, 6 or even 9 months, during which economies are expected to suffer dearly. A worse case scenario is a half- or full-year of partial lockdown which could result in a minimum 10-20% in aggregate negative economic impact.

  • (4) Cross border trade impacts are harder to model reliably without more information on trade partners and networks for each country, as well as the nature of the products and services being imported or exported. However, assuming (rather carelessly!) an equivalent impact on both sides of imports and exports, it is the net exports that matter. Referring to the earlier table on net export contribution to GDP, one can estimate this impact to be limited to <5% of GDP assuming the bulk share of negative trade impact can be contained below 6 months. This is an assumption and not at all a certainty. Mismatched timing of recovery between export and import trade could yield disastrous consequences if not managed well.

  • So where does this leave us? In short, a country with a prolonged struggle restricting mobility, and heavily reliant on trade and tourism could be looking at an annualised negative impact of some 15-25% of their historic GDP levels, whereas countries on the other extreme could be looking at a contained ~5-10% impact. As such, we believe economic forecasts are underestimating the impact of the pandemic to at least some if not a great many number of economies.

  • Note: this computation blatantly excludes the impact of emergency stimulus packages that intend to temporarily buoy economies through the recovery process - see figures below for some indicative sizes. While we will not deliberate them here, a cursory look suggests the amounts have been significant, and may well be very meaningful to bounce economies back to baseline trajectories.

Update 7 June: a follow up to this article is available here.

Figures: Global stimulus packages announced by governments. Source: Statista


Recent Posts

See All