The long decline

The COVID-19 pandemic provided the world economy with its biggest jolt in living memory. According to the World Bank, output shrank 3.6 per cent in constant US dollar terms relative to 2019 - the deepest annual contraction since the Second World War.

The recovery is poised to be just as spectacular. The International Monetary Fund recently predicted output was on course to rebound six per cent in 2021 and 4.9 per cent in 2022. Somewhat improbably, the US economy is now forecast to be bigger in 2023 than it would have been had COVID-19 never happened, thanks in large part to huge government handouts.

However, while the worst of the pandemic appears to be behind most of the developed market economies, aided by the ongoing effectiveness of several vaccines, their longer-term prospects appear less rosy.

Across the developed world, economic growth rates have been trending steadily lower for more than half a century, as shown in Figure 1. The same phenomenon has begun to affect some important developing nations, such as China and South Korea too. Once the tumult created by the pandemic begins to subside, it seems likely the forces that precipitated this secular decline in trend growth will quickly reassert themselves.

Figure 1: Secular decline in G7 trend GDP (per cent)

Note: Shows annualised change in GDP over rolling 5-year periods, constant 2010 US dollars.
Source: The World Bank, Aviva Investors’ calculations, October 2021

Demographic disadvantage

To understand the drivers of economic growth, and how trends can change over time, it is useful to decompose national income into its constituent parts.

Y = National income
Y/H = Productivity
H/E = Average hours worked
E/N = Employment rate (= 100 – unemployment rate)
N/P_wa = Participation rate
P_wa = Population of working age

Essentially, there are two ways to grow an economy: either the number of hours worked increases or labour productivity rises. On both counts, many of the world’s biggest economies face stiff headwinds.

Over the long run, a country’s working-age population tends to be the most important driver of the number of hours worked. Using the demographic projections of the Organisation for Economic Co-operation and Development (OECD) out until 2050 – which essentially assume recent trends with regards to birth and death rates and migration continue – the main conclusion is clear and stark.

In the next three decades, ageing populations are set to produce massive challenges for most developed, and some important emerging countries. In almost all cases, and all the economically important ones, the working age population (20-64) is set to grow much more slowly than in the past, and in most cases will shrink. This is already happening in many countries.

Taking the OECD’s forecasts for seven advanced countries – the US, Japan, Germany, UK, France, Italy, and Spain – and two important developing nations – China and South Korea – we see the aggregate working-age population has already started to fall and the rate of decline is expected to accelerate. The same trend holds when stripping out the impact of an expected decline in the population of working age in China, by far the largest of the nine.

Figure 2: Population growth goes into reverse (in mn)

Source: OECD, October 2021

Accounting for nearly two-thirds of global output in 2020, these nine countries serve as a reasonable proxy for the world economy. As such, the headwinds to potential world GDP growth from demographic trends alone over the next 30 years look daunting.

It is possible to get a better sense of the magnitude of the problem by considering how these trends translate into annual growth rates. The chart below reveals that, whereas the average growth rate of the population of working age for this nine-country aggregate was 1.38 per cent a year between 1960 and 2020, and 0.94 per cent over the last three decades, over the next 30 years it is set to be -0.54 per cent.

Figure 3: Working-age population, annual per cent change

Source: OECD, Aviva Investors’ calculations, October 2021

Without an increase in average hours worked, higher employment, rising labour force participation rates, or a pick-up in productivity, this swing of nearly two percentage points means annual growth in world GDP, as proxied here by these nine countries, will be nearly two percentage points lower a year, on average, over the next three decades than over the last six.

Crude averages such as this mask big differences between countries. To contemplate the scale of the step change facing some, consider for instance that in 2016 there were 37 million Italians of working age and 935 million Chinese. If the OECD’s projections are right, by 2050 those figures will be 29 million and 770 million respectively – falls of around 17.5 per cent. Korea and Japan are expected to experience even bigger drops – 34 per cent and 30 per cent respectively, while the working-age population is expected to shrink in Germany by 12 per cent, in France by five per cent and in Spain by nine per cent.

According to the OECD’s forecasts, only in the UK and US is the working-age population set to grow, by two and 12 per cent respectively. This is largely because both have had higher birth rates and more migration than the other seven countries in the recent past. While these trends are projected to continue, there is a danger both countries’ positions will begin to look far less healthy thanks to tougher stances on immigration.

Although demographic trends can reverse, changes take a long time to take effect. For example, a higher birth rate could help offset current trends, but it would be 20 years before the impact on the working-age population were felt. In the meantime, labour participation may dip as young dependents are looked after by their parents. Immigration could increase to provide another offset, but there are attendant issues regarding attitudes to migrants and problems of acceptance and assimilation. Globally, it also represents a zero-sum game.

Flexible working and increased participation?

Although one of the other four determinants of economic growth could offset the impact of the worsening demographic picture, current trends do not give much cause for hope. Take average hours worked: across the world, this has been in decline for 60 years.

One of the most significant changes to labour market practices in recent years has been the move towards flexible working. Part-time workers now account for a much bigger share of the workforce. But it has not just been the growth of part-time employment that has depressed average hours worked. Even those in full-time employment have been working steadily less hours as more opt for leisure over work. Figure 4, which shows only the broad aggregate, reflects both effects.

Figure 4: Average weekly hours worked per worker (OECD countries)

Source: OECD, Aviva Investors’ calculations, October 2021

Even if 2020’s sharp COVID-related drop may be about to reverse, countries are unlikely to get any boost to growth from increases in the average numbers of hours worked per worker – unless habits change.

Instead, it appears more likely many nations will see average hours fall as they follow the normal evolution. Any moves towards even greater leisure time being taken in the future – mooted by many as a realistic probability – would mean an additional negative kicker for growth.

While Figure 4 suggests the average working week may settle close to 33 hours for all the countries analysed by the OECD, there is little reason to expect a reversal of this trend. No further decline in average hours worked appears the best that can be hoped for.

There may be scope for some increase in labour participation rates, which has been on a rising trend in most countries for many decades, as illustrated in Figure 5.

Figure 5: Labour force participation rates

Source: OECD, October 2021

This has been an important driver of GDP growth for more than half a century, and largely reflects rising female participation rates. However, while some countries may have scope to lift female participation rates further, it is unclear all do. Furthermore, any further increase is likely to happen gradually.

Another notable trend has been a steady rise in the number of elderly people staying in the workforce. In several countries, participation has been increasing fastest in older age groups. If people can be encouraged to stay in work for longer – or if they feel they must, due to the inadequacy of their pensions, or changes in pensions legislation – labour supply will be higher than it would otherwise have been.

However, while there may be no reason to expect rising participation to tail off in the near term, it cannot continue at the same pace indefinitely. In any case, although greater participation rates may offset some of the impact from lower population growth and reduced average working hours, on average they could only realistically boost GDP growth by 0.2 to 0.3 percentage points a year, and even that may be optimistic.

As for employment rates, countries with a relatively high number of people out of work may be able to boost growth in the short term by absorbing more people into productive employment. Even if this were achieved, it would be a one-off adjustment and would fail to significantly impact trend economic growth. For a more meaningful long-term contribution, labour market reforms would be needed. History suggests they are notoriously hard to push through, however, and it would be unwise to bank on them.

Can higher productivity come to the rescue?

For the long-running decline in trend economic growth in some of the world’s biggest economies to be arrested, much would seem to depend on what happens to productivity. Again, there does not appear to be much cause for optimism and it would require a spectacular reversal of a long-run trend.

For the past 60 years, productivity growth has been in a secular decline. This is to be expected and universally accepted as a ‘normal’ economic trend. In the early stages of development, productivity gains are easy to achieve, but as economies mature, incremental gains become more difficult to eke out. This is exacerbated by the transition away from manufacturing towards services. Productivity growth is always stronger in the former, as it is generally possible to increase the capital stock to boost output per hour worked. Within the service sector, although the same principle applies, more output often requires greater human capital too.

The experience since the Global Financial Crisis has been particularly worrisome, with productivity growth very low by historical standards around the world. Although it is possible robotics and artificial intelligence will reverse this trend, it seems equally valid to ask whether the world has entered or is entering a new low-productivity-growth regime.

It is possible the persistence of ultra-loose monetary policy is exacerbating the problem. Far from incentivising productive investment, it could be discouraging it. Public and private companies have taken advantage of low interest rates to issue record amounts of debt over the past decade. But since an immediate financial gain today is worth more than a potential future one through investment, rather than being used to fund new capital projects, companies have all too often found it more profitable to focus on financial engineering instead.

Figure 6: Gross fixed capital formation - high income countries (per cent of GDP)

Source: The World Bank, October 2021

What this means for growth and asset prices

By combining the five right-hand side variables in the calculation of national income and adopting sensible assumptions for the likely time path for all of them, we can produce GDP projections for individual countries over the longer term. We assume:

1. The working age population follows the OECD’s demographic projections.
2. Average hours worked stabilise at current levels.
3. The upper limit for labour participation rates is 80 per cent and countries converge towards this rate over the long run.
4. Employment rates converge towards our estimate of each country’s natural rate of unemployment over the next four years.
5. Productivity growth is unchanged versus the past decade.

Figure 7 shows both past GDP growth rates for the nine countries included in this analysis and our estimated projections for the future based on the above assumptions. The differences are stark and make for disturbing reading. We estimate there is likely to be a reduction in trend GDP growth of between 0.5 and one percentage points a year – significantly more in some cases – for multiple years.

Figure 7: Average annual GDP growth (compounded rate, constant local currency terms)

Source: The World Bank, Aviva Investors’ forecasts, October 2021

Taking the US as an example, we find that whereas GDP grew at an average annual rate of 2.4 per cent between 1995 and 2020, it is likely to average closer to 1.6 per cent over the next three decades. The decomposition of these two growth rates is shown in Figure 8.

Figure 8: US – decomposition of GDP growth

Source: The World Bank, Aviva Investors' calculation, October 2021

It is hard to resist the conclusion GDP growth rates will be lower, and sometimes much lower, than historical levels. Changes of this magnitude have not been felt for centuries. Anything that depends on, or is related to, the underlying trend rate of GDP growth will be significantly impacted. Households, companies, and governments will have to adapt to a very different environment. It is fair to assume most are unprepared.

This will also have major ramifications for financial markets. The paths for GDP and inflation are critical influences on the prices and valuations of all financial assets. Yields on sovereign bonds have historically been closely correlated with nominal GDP growth, while the income streams from real assets such as property and equities are typically directly related to nominal GDP as well.

In simple terms, if nominal GDP rises, the equilibrium level of rents or earnings or dividends tends to rise at a comparable pace. Forward-looking investors will need to keep a close on these underlying drivers of economic growth.