Market Perspective – Recessions and portfolios: It all depends
Kevin Gardiner, Global Investment Strategist, Wealth Management
Unloved cycle enters new territory
This may shortly become the longest-ever US expansion. If it lasts until mid-year, it will nudge the cycle of 1991-2001 off the top spot in the National Bureau of Economic Research's (NBER) league table dating back to 1860.
Some investors are nervous - its great age seems to suggest the end is overdue. Should they prepare for its eventual demise - and if so, how?
We have given it the benefit of the doubt so far. But many pundits wrote it off from the start - it has been the least loved of any recent cycles. Lots of 'clever money' has been selling or ignoring it all the way up.
It's considered smart to say the US can't grow much any more (“too much debt”, “secular stagnation” and all that). In fact, growth in real household and business spending has averaged 3.1% - little different to the 3.4% in the previous upswing. This ought not to have been a surprise.
It's also fashionable to say the US doesn't matter so much any more. But while China may have contributed more to global GDP gains recently, the growth that matters most to capital markets is (still) in the US. And because US imports are bigger and grow faster than its exports, it has made a bigger international contribution.
Nor have the gains been quite as uneven as suggested. The unemployment rate is the lowest since the 1960s, and widely quoted median household income data can be misleading. It's a little inconsistent to believe, as well-meaning commentators do, both that quantitative easing wealth effects have driven the upswing and that growth has been unusually unequal - wealthy consumers don't account for much of total spending. And contrary to what pundits seem to suggest, there has been no golden age of equality.
Meanwhile, our main reason for giving the US cycle the benefit of the doubt remains intact: the absence of excess. Consumers have not been borrowing - or banks lending - recklessly. Inflation remains subdued.
Nonetheless, underappreciated and well-behaved though it is, the US upswing will end at some stage, and the economy will shift into reverse gear for a while. The likely human costs are clear: a rebound in unemployment and a fall in incomes. But what might it mean for investments?
The frustrating but honest answer is: it all depends. We often cut corners, and use the term 'recession' indiscriminately to signify A Bad Thing. But downturns vary tremendously - and their impact on portfolios varies even more.
Spoiler alert: forecasters beware
The end of an NBER expansion is not the only definition of recession. The NBER dates the US cycle to the nearest month, using several indicators, whereas a recession is generally defined - anywhere - as two consecutive quarterly falls in GDP. But both denote shrinking economies.
We may not spot the next US contraction beforehand. We will do our best, but the exact timings of most recessions are unpredicted. Most of the recessions that are predicted don't happen, but turn out to be the mistakes of more routinely unhappy forecasters.
Why are contractions so difficult to spot? They are quite rare: most of the time, economies grow. Unexpected events - such as 9/11, or the seizure of the global money supply in 2008 - can knock economies off course. But usually, downturns arrive unexpectedly because the economy is driven by people, people are driven by emotion, and emotions fluctuate.
Confidence can swing sharply and infectiously. Knowing when our collective mood will shift away from seeing discretionary purchases (the new car or an upgraded production line) as life and business-enhancing assets, and see them instead as liquidity-draining burdens, is an art, not a science. And it is an art without any great artists.
Can't central banks abolish recessions?
In economists' ideal world, the instant a downturn seemed imminent, interest rates would be cut, and/or public spending raised and/or taxes reduced, neatly cancelling it. Similarly, if it looked as if growth were too strong for comfort, the authorities could raise rates and/or tighten fiscal policy.
This 'control engineering' view was first popularised in the 1950s. But in practice, instead of smoothing the cycle away, the lags in the system - it took time to identify the problem, to decide what to do, and then to act - often resulted in the cycle being amplified, not smoothed. By the time the economy was given a boost, for example, it had begun to revive of its own accord. This was the 'stop-go' era.
More timely data, and the success of counter-cyclical policies in 2008/9, have encouraged economists and central bankers to think such cyclical fine tuning is once again worth trying. We are sceptical. 'Control' problems remain, and a completely stable world might not be the boon economists imagine.
It can still take some time to decide that an economy may be on the turn: a single month's data is rarely enough. In 2008, Lehman Brothers' collapse followed a prolonged period of market volatility during which bankruptcy did not seem inevitable (in contrast to the perfect view accorded to hindsight in The Big Short).
Central banks hope that new machine-learning techniques and big cross-sectional datasets will help spot major downturns quicker. We doubt any patterns identified will be meaningful, and there are too few events to allow us to talk of a 'typical' crisis.
Control remains difficult too because the links between interest rates, for example, and the economy are looser than economists imagine. Talk of 'transmission mechanisms' is misleading. There are just too many moving parts, and they have minds of their own.
A Truman Show economy
Perfect control might in any case prove counterproductive. A world without economic accidents would be a bit like The Truman Show. Truman, “on air, unaware” as the subject of a reality TV programme, is protected from all sorts of mishaps - including the use of his own free will.
It might seem a nice problem to have. Truman's world is well-ordered and superficially cheerful. But we'd lose something important - just as Truman, cocooned against misfortune, is missing out on real life.
A world seemingly without cyclical risk might be poorer in the longer term. It would lack urgency and spontaneity. Risky assets might be repriced - perhaps mistakenly. Eliminating dangers posed by a changing business climate - the exuberant overinvestment that follows (for example) technological breakthroughs, or the discovery of new resources, or free cashflow - might also eliminate many opportunities.
We need adequate insurance for those displaced by the cycle. And there are of course limits to the risks we might embrace in the spirit of creative destruction - the authorities should strive to avoid major crises and systemic collapse. But if the real world is unavoidably and profoundly uncertain, carefully designed artificial ones may have less appeal. We don't share the growing belief that, having saved the world in 2008, central banks should try to abolish the cycle completely. It smacks of hubris.
This doesn't mean we're looking forward to the next recession. But a downturn sooner or later is likely anyway (those control problems), and can be seen as part and parcel of the workings of the market economy - the least bad economic system we've yet discovered.
Theory and past practice
Recessions affect stocks mostly via their impact on business: falling revenue, squeezed margins, enhanced credit risk, asset writedowns. We'd expect their portfolio impacts to vary. Are companies and/or the banking system especially fragile, and prone to contagious systemic risk? Is it a complete surprise, or partly priced in?
And, importantly, how will the authorities respond? Stock prices can be affected by interest rates, and if the latter fall far enough - whatever our misgivings about the possibility/desirability of perfect control - stocks might even respond positively to the downturn.
History confirms there is no such thing as a typical recession or market response. Figure 1 shows NBER downturns, and the associated performance of GDP, corporate earnings and the stock market. The falls in GDP look small, but remember GDP is a measure of economy-wide value-added growth, and much less volatile than corporate earnings - a drop of 'just' 3% in GDP is actually a big cyclical deal.
Click the image to enlarge
Clearly, it would have paid to have avoided most downturns. But in some cases, by the time the event occurs and portfolio decisions have been taken and implemented, the picture is less clear.
On 'average', stock markets have peaked eight months before an NBER recession starts, and are broadly unchanged eight months after it. Even though its exact timing may still come as a surprise to economists, when it does start it thus turns out to have been effectively partly priced in (Figure 2).
Click the image to enlarge
The 1990 recession followed a prolonged period of economic excess, but the market peaked before it started, and rallied before it ended. You could have ended up selling after the fall and missing the rebound. There have been other instances too where the total drawdown was brief, or just not alarming.
One of the biggest market falls was associated with only a modest economic dip, in 2000. In current data it no longer qualifies as a conventional GDP recession. It was the bursting of the 'new economy' bubble - a reminder that the stock market is quite capable of starting a fight in an empty room, as it were. The most recent market drawdown - the GFC in 2008/9 - was certainly associated with a big economic hit. But again, the trauma really originated in the markets, not the economy.
If we do get wind of a looming recession, and think it could be big enough - or not priced in enough - to try to avoid, what would we advise?
A carefully designed portfolio will hold some diversifying assets alongside procyclical ones. Figure 3 focuses on periods in which GDP was falling, and shows, as we'd expect, that bonds and defensive sectors usually do best. But it is not easy: while bonds did better than stocks in the inflationary 1970s recessions, they still delivered negative returns (after inflation). When GDP fell in 1980, stocks actually beat bonds, and delivered positive real returns. And a sector's cyclical sensitivity can vary with its valuation, balance sheet leverage, interest rates - and its shifting footprint.
Click the image to enlarge
Are 'defensive' consumer staples, for example, necessarily safer when the value of their brands - and distribution networks - is being questioned?
Conclusion: keep an open mind
We do not yet think the next US recession is imminent. The 'usual suspects' that have triggered downturns in the past - inflation, reckless borrowing/lending, market bubbles - are missing. This is no guarantee, but we continue to give the cycle the benefit of the doubt.
There will be one: recessions are unavoidable. But their impacts vary hugely, and a world without them could be a poorer one.
After the GFC, many pundits assert that the next economic/market setback must be equally seismic. It could be - but there is no reason it has to be. It may turn out - eventually - to have more in common with 1990 than with 2008.
Meanwhile, remember that net of recessions and drawdowns, in inflation-adjusted terms the US economy is nine times bigger than in 1950, and the stock market's total return index is 146 times as high. As long-term investors, we shouldn't let these signals be obscured by the cyclical noise, however distracting it is at the time.
Now, back to watching the numbers…
In this Market Perspective:
Download the full Market Perspective in PDF format (977 KB)
Listen to the latest Market Perspective podcast