Global Investment Outlook 2020 — The Pandemic Year

2020 is remembered as the pandemic year, but for allocators its significance is different from its public memory. Within eleven months it contained a collapse of extraordinary speed, an equally extraordinary recovery, and a policy intervention large enough to change how a generation of investors understands central bank behaviour.

The initial decline in late February and March was not a considered repricing of corporate earnings. It was a scramble for cash. Correlations across nearly all assets converged towards one — the specific condition in which diversification stops working, because the thing being sold is not any particular risk but illiquidity itself.

What followed was the more instructive half. Markets began recovering while the epidemiological and economic news was still deteriorating. This looked irrational to many observers and was widely described as a disconnect between Wall Street and Main Street. It was neither irrational nor a disconnect. Central banks cut policy rates to near zero and committed to asset purchases at unprecedented scale, including — in the United States — corporate credit. That simultaneously collapsed the discount rate applied to future cash flows and removed the tail scenario in which otherwise-viable companies failed for want of financing.

An asset's price reflects future cash flows discounted to today. 2020 damaged the near-term cash flows badly. It reduced the discount rate applied to all of them more.

The March liquidity event

The first phase of the drawdown behaved unlike a normal bear market, and the difference matters.

In a conventional decline, capital rotates: equities fall, government bonds rise, gold typically holds. In March 2020, for a period of roughly two weeks, almost everything fell simultaneously. Treasuries — the asset investors buy precisely when frightened — sold off. Gold fell. Investment-grade credit fell alongside high yield.

This is the signature of a liquidity event rather than a valuation event. When institutions face margin calls, redemptions, or an urgent need for cash, they cannot sell only the assets they consider overvalued. They sell what has a buyer. In a stressed market, that means selling the highest-quality, most liquid holdings — which is why safe assets fall hardest at the exact moment their safety is most wanted.

Diversification protects against assets falling for different reasons. It does not protect when the reason for selling is that cash is needed and the asset happens to be sellable.

The intervention that ended this phase was not primarily a stimulus for the economy. It was a commitment to act as buyer of last resort, which restored confidence that a seller could find a bid. Once that was established, the forced-selling dynamic unwound quickly.

Why markets recovered before the news did

The apparent paradox of 2020 — sharp recession, strong asset returns — dissolves once the arithmetic is stated explicitly.

An asset is worth its future cash flows, discounted. Two things changed in 2020, in opposite directions:

Near-term cash flows fell sharply. Revenues collapsed in travel, hospitality, physical retail and energy. For most affected businesses this was a hit concentrated in roughly four to eight quarters.

The discount rate applied to all future cash flows fell to near zero, and was expected to stay there for years. This lifted the present value of every cash flow arriving after the disruption — which, for a going concern, is the overwhelming majority of its value.

For a business expected to operate for decades, two bad years are a small fraction of total value. A durable reduction in the discount rate is not. The second effect outweighed the first, and asset prices rose while GDP fell.

This also explains the composition of the recovery. It was concentrated in long-duration assets — technology, software, growth businesses whose value sits largely in distant cash flows — precisely because those are the assets most sensitive to a change in the discount rate. The same mechanism, running in reverse, would drive the 2022 correction.

Private markets: a pause, not a repricing

Private capital behaved very differently from public markets, and the difference is instructive about how private marks work.

Activity essentially stopped for a period in the second quarter. Diligence requiring travel or site visits was impractical. More fundamentally, nobody could price anything: valuation depends on a forward view, and in April 2020 the range of plausible outcomes was extremely wide.

What did not happen was a large, durable markdown. Some managers wrote down positions at the first quarter mark, but many of those write-downs were reversed within two quarters as public comparables recovered.

The effect on the reported record is significant. Looking only at annual private market returns, 2020 appears to have been a good year with little disruption. That impression is an artefact of quarterly appraisal-based marking. The disruption was real; it simply occurred and reversed between observation points.

This is the same smoothing mechanism that would matter enormously in 2022 — but in 2020 it flattered private marks, whereas in 2022 it delayed necessary declines. The mechanism is neutral; only its direction changes.

The pull-forward problem

The most consequential analytical question of 2020 could not be answered in 2020, and getting it wrong caused a great deal of subsequent misallocation.

Lockdowns produced enormous, immediate adoption of digital services: video conferencing, e-commerce, streaming, remote collaboration, telehealth. Growth rates in these categories reached levels that would ordinarily take years.

The question was whether that represented a permanent step change in behaviour or a temporary pull-forward of demand that would otherwise have arrived gradually.

The distinction is not academic. If adoption is a step change, the new growth rate is the baseline and valuations should reset upward. If it is a pull-forward, the elevated growth is borrowed from future periods, and growth should decelerate below trend once conditions normalise.

Both readings were defensible in 2020 and the data could not distinguish them, because there was no counterfactual. Capital was allocated overwhelmingly on the step-change reading. In hindsight the truth varied enormously by category — some behaviour changes proved durable, others reverted almost entirely — but the uniform application of the optimistic reading is the root of a great deal of 2021 overvaluation and 2022 disappointment.

The general lesson is worth stating plainly: when a shock produces a sudden change in a metric, the critical question is whether the shock changed the underlying preference or merely the timing of its expression. That question is usually unanswerable while the shock is ongoing, which argues for holding both readings rather than committing to the more attractive one.

Sector dispersion was extreme

2020's aggregate figures conceal a dispersion between sectors wider than in almost any comparable year.

Severely impaired: aviation, hospitality, physical retail, commercial real estate, energy. These faced not reduced demand but, in several cases, demand approaching zero for a period — a condition few business models are built to survive.

Structurally advantaged: software and cloud infrastructure, e-commerce and logistics, digital payments, biotech and life sciences, semiconductors. Some benefited from accelerated adoption; others, particularly life sciences, from a step change in public and private funding for the sector.

Genuinely ambiguous: commercial real estate and office in particular, where the question of long-run occupancy would take years to resolve and remains partially open.

For allocators the practical consequence was that index-level exposure meant something quite different in 2020 than usual. A market-cap-weighted index concentrated exposure in exactly the long-duration technology businesses most favoured by falling discount rates — which produced strong returns and, simultaneously, an unhedged and largely unexamined bet on rates staying low.

What the March liquidity event taught about liquidity

March 2020 is the most instructive liquidity episode of the period covered by this archive, and its lesson is frequently reduced to "liquidity matters" — which is true and not useful.

The specific finding was that liquidity is a property of conditions, not of an asset. An asset that trades continuously in normal markets is not therefore liquid. Its liquidity depends on the presence of a counterparty willing to take the other side, and that presence is itself conditional.

What failed in March was not the assets but the intermediation. Dealers who normally provide two-way markets reduced their willingness to hold inventory, because their own risk limits tightened as volatility rose. That is the same mechanism the 2017 and 2018 reports describe operating in position sizing: a risk model responding to observed volatility reduces capacity at exactly the moment capacity is most needed.

The consequences ran in a specific order, and the order is the useful part:

• The most liquid instruments came under pressure first, because they are what a seller reaches for when they need cash. An investor needing liquidity sells what they can sell, not what they most want to sell.
• That produced the counterintuitive pattern of high-quality government bonds — the safest assets — experiencing dislocation, while less liquid assets showed no visible stress simply because nobody was transacting in them at all.
• Absence of price movement was mistaken for absence of stress. An asset with no trades has no price, which is a different condition from a stable price.

The practical implication is that liquidity planning must be done in advance and must assume the plan will be tested. An investor whose contingency is "we will sell the liquid portion" should check what the liquid portion is worth when everyone else has the same contingency. The answer in March 2020 was: less than expected, and available only because policy intervened.

A portfolio's real risk is not only what it holds. It is whether it can transact when it must — and that is a question about who will be on the other side, not about what the asset is.

What an allocator could act on

2020's mechanisms translated into specific, checkable positions.

Identify what the portfolio's actual macro bet is. Many 2020 portfolios that looked diversified by sector were concentrated on a single variable: the discount rate. A cap-weighted equity index in 2020 was substantially a long-duration position. Diversification by label is not diversification by driver, and the check is to ask what single variable, if it moved, would move most of the portfolio.

Separate the pull-forward question from the growth question. A category experiencing abrupt adoption may be seeing a permanent level shift, a permanent rate shift, or demand brought forward. These have very different valuation implications and were routinely conflated. The distinguishing evidence is what happens to growth after conditions normalise, which means the question cannot be settled at the time — but it can be held open rather than resolved by assumption in the most favourable direction.

Do not extrapolate from the steepest part of a curve. Annualising a peak-period growth rate implicitly assumes the steepest observed growth is the new run rate. It is the single worst point on the curve from which to extrapolate, and it was the point most 2020 valuations were set from.

Treat the absence of a markdown as absence of measurement. Private portfolios showed limited drawdown in Q2 2020. That reflected quarterly appraisal-based valuation, not resilience. A position that has not been remeasured has not held its value; it has not been asked.

Recognise that a demonstrated policy reaction function changes behaviour, and that behaviour has a precondition. Investors reasonably concluded that authorities would intervene at scale to prevent financial-system failure. That conclusion rested on inflation remaining below target, which gave policymakers room. The 2019 report makes the same point about the pre-2020 asymmetry belief, and the precondition failed for both in 2021.

What 2020 established

Several things emerged from 2020 that shaped the following years.

• The policy reaction function was demonstrated. Investors observed that authorities would intervene at extraordinary scale and speed to prevent a financial-system failure. That belief altered risk-taking for years afterwards.
• Long-duration assets were revealed as a rates bet. Many portfolios that appeared diversified by sector were in fact concentrated on a single macro variable.
• Liquidity was re-established as a first-order risk. March demonstrated that a portfolio's real risk is not only what it holds but whether it can transact when it must.
• The measurement gap in private markets became visible to anyone comparing the lived experience of Q2 2020 with the reported annual return.