1. General assessment of the macroeconomic situation

Global activity is now above pre-pandemic levels, but the recovery is incomplete

The recovery in global economic activity since mid-2020 has been more vigorous than expected (Figure 1.1), with output in most OECD countries is now close to or above pre-pandemic levels. This reflects the prompt and massive policy support for firms and households from the outset of the crisis, including the additional measures announced this year, successful public health measures to limit transmission of the COVID-19 virus and, above all, the rapid rollout of effective vaccines.

The strength of the rebound has not yet permitted a full healing of the global economy from the effects of the pandemic. The world has foregone the growth that would have occurred in 2020 and the first half of 2021: global GDP in mid-2021 was still 3½ per cent lower than projected before the pandemic. Moreover, this foregone growth to date has not been distributed equally: the loss has been proportionately greater for middle-income emerging-market economies as a group than for advanced economies, and greatest of all for low-income developing countries. Contact-intensive sectors and lower-income households have also been particularly hard-hit, contributing to the incomplete recovery in labour markets. Among OECD countries, about 7½ million fewer people were in work in the third quarter of 2021 than in the fourth quarter of 2019. Many emerging and developing economies have also suffered declines in employment during the pandemic (ILO, 2021) and poverty has risen. Even in economies where the number of people in work by mid-2021 was close to or even above its pre-pandemic level, total hours worked were often still lower than in late 2019 (Figure 1.2).

Figure 1.1. The global recovery has outstripped expectations but is not complete

Source: OECD Economic Outlook 110 database; OECD Economic Outlook 106 database; OECD Economic Outlook 107 database; OECD Economic Outlook 108 database; OECD Economic Outlook 109 database; and OECD calculations.

 StatLink https://stat.link/9vr0w2

Figure 1.2. The labour market recovery is only partial in most countries

Percentage change between 2019Q4 and 2021Q2

Note: National accounts based measure of employment for all EU member states, Norway and the United States, and labour force survey measure for remaining countries.

Source: OECD Economic Outlook 110 database; Bureau of Economic Analysis; Statistics Canada; Australian Bureau of Statistics; Statistics Bureau, Japan; Eurostat; Office for National Statistics; and OECD calculations.

 StatLink https://stat.link/8ir5zk

Quarterly global GDP growth is estimated to have remained broadly stable in the third quarter of 2021, with softer outcomes in China and the United States offset by steady growth in the euro area and a rebound in India and some energy-exporting economies. Third quarter output was particularly weak in a number of Asia-Pacific economies, including Japan and Indonesia, reflecting the extent to which activity has been hampered by the Delta variant and supply constraints. Recent high-frequency indicators generally suggest that the pace of the recovery may have eased (Figure 1.3). Location-based indicators of mobility have continued to improve in emerging-market economies, but have begun to ease in the advanced economies, especially in some European countries. Business survey measures of output and new orders are below their peaks earlier this year, and consumer confidence has slipped back in recent months in many countries. Global industrial production and consumer retail spending are above their corresponding levels in 2019, but momentum has eased. The car sector has been particularly affected by supply bottlenecks, and output and sales have fallen sharply in recent months (Box 1.1), holding back the overall recovery.

Figure 1.3. The momentum of the recovery has eased

Note: Data in Panels A, C and D are PPP-weighted aggregates. The retail sales measure uses monthly household consumption for the United States and the monthly synthetic consumption indicator for Japan. Data for November in Panel A are based on information up to November 21.

Source: Google LLC, Google COVID-19 Community Mobility Reports, https://www.google.com/covid19/mobility; Markit; OECD Main Economic Indicators database; Refinitiv; and OECD calculations.

 StatLink https://stat.link/od16bj

Box 1.1. The impact of supply-side disruptions on the automotive sector

The car sector is being affected severely by supply-chain disruptions

Shortages of semiconductors and other intermediate goods, delays in supplier delivery times and bottlenecks in shipping have hit production in many industries. Recent car models typically contain more than a thousand chips governing a range of essential operations (e.g. engine cooling, battery management, tyre-pressure monitoring, interior lighting, and emergency braking). As a result, the car sector has been especially affected by disruptions, with large production cuts despite strong global demand.

The automotive sector is of particular interest because it plays a major role in the manufacturing sector in many countries, and accounts for a sizeable share of GDP. Prior to the pandemic, motor vehicle production accounted for 2% of GDP in the euro area, 4.7% in Germany, 0.75% in the United States and 3% in Japan. In the first nine months of 2021, motor vehicle production in the euro area was 26% lower than in the same period of 2019, 30% lower in Germany, 10% in United States and 23% in Japan (Figure 1.4, Panel A).

The reduced availability of new cars and low inventory levels have also hit global car sales, which declined by 40% in both Japan and the United States between April and September, and by 16% in Germany (Figure 1.4, Panel B). Global car sales declined by over 20% over the same period, a pace of decline only seen previously in deep recessions, such as the early stage of the pandemic last year. The combination of solid demand and the reduction in supply has placed upward pressure on prices. New and used car prices have risen substantially this year in several countries, especially the United States (Helper and Soltas, 2021).

Figure 1.4. Car production and sales are both contracting

Source: Eurostat; and Refinitiv.

 StatLink https://stat.link/3m2eyc

Given the large share of trade-related output throughout the transport equipment sector,1 recent supply disruptions have played a critical role in the sizeable sectoral decline observed around the world. The output of the transport equipment industry is strongly linked through global value chains, especially in Europe. In 2019, more than 80% of the production in Hungary, Slovakia and Slovenia involved goods that crossed at least two borders along the production process, with the share above 50% for most European countries.

A more detailed, though less up to date, breakdown of transport equipment subsectors is available for 2015. This shows that the motor vehicle industry mostly consumes inputs from Europe (53% of inputs) and Asia (22%). Globally, electronics and electrical equipment represent 6% of total inputs in the sector. Automotive manufacturers typically source these inputs from trade partners in the same region (Figure 1.5). In Japan and Korea, around 90% of the electronics and electrical equipment inputs in the transport sector come from Asia. The United States and Canada also mostly rely on regional value chains, with 65-75% of electronics and electrical equipment inputs sourced from the Americas.2 In many central and eastern European countries, the share of inputs from Europe is above 70%.

Figure 1.5. Regional value chains are important in the motor vehicle sector

Value share of electronic and electrical equipment inputs from different regions

Note: America includes Brazil, Canada, Mexico and United States. Asia and Pacific includes Australia, China, Chinese Taipei, Indonesia, India, Japan and Korea. Based on input data in current prices and expressed in USD terms.

Source: World Input-Output Database (WIOD), 2016 release.

 StatLink https://stat.link/yv0o3x

What is the contribution of supply-side disruptions to the contraction of motor vehicle production?

The separate contributions of supply and demand imbalances to the decline in motor vehicle production can be assessed empirically. Econometric estimates for a group of 30 OECD and emerging-market economies show that PMI indicators of suppliers’ delivery times and new orders (a proxy for demand) both have significant effects on motor vehicle production. The estimated relationships can be used to derive the difference between car production in 2021 and the output level that might otherwise have been expected from demand growth if suppliers’ delivery times had remained unchanged from the level in the fourth quarter of 2020. This gap is substantial in a number of countries, implying that higher supplier delivery times and other disruptions may have reduced motor vehicle production in Germany by an amount equivalent to over 1½ per cent of GDP in the first nine months of 2021 (Figure 1.6, Panel A), and production in the Czech Republic, Japan and Mexico by an amount equivalent to between ½ and 1% of GDP. In many of the other major producers of motor vehicles, the impact of supply constraints on production is smaller, and motor vehicle production represents a lower share of overall activity, but there is still a noticeable drag on GDP this year. Aggregating across countries, there is a growing gap between actual motor vehicle production and what might have occurred in the absence of supply constraints (Figure 1.6, Panel B), with actual production around 25% lower this year than the scenario with no additional supply constraints in 2021.

Figure 1.6. Supply-side constraints are depressing car production in 2021 and weighing on GDP

Note: Panel A shows the implied impact on GDP from the gap between actual motor vehicle production in the first nine months of 2021 and production in a scenario in which new orders (a demand proxy) follow their actual path in 2021, while suppliers’ delivery times are held at their 2020Q4 level. Panel B shows the overall gap when aggregated across countries between production in the scenario and actual production, for countries accounting for 85% of global motor vehicle production. Based on estimates for a panel of 30 OECD and emerging economies, over April 1996 to December 2019, with monthly growth rates of car production regressed on country-specific PMI indices, lags of the dependent variable and country and time fixed effects. PMI indices for the overall manufacturing sector are weighted by the share of motor vehicles output in total manufacturing output.

Source: Markit; Eurostat; OECD Structural Analysis database; OECD Trade in Value Added database; and OECD calculations.

 StatLink https://stat.link/qpgu97

Supply-side strains are expected to persist

At the onset of the pandemic, when automotive producers cut orders for chips in anticipation of lower demand, semiconductor producers shifted part of their production to supply producers of computers, webcams, tablets and other electronic and communication equipment, where sales were soaring. This redirection of semiconductor supply has contributed to the shortages carmakers are facing currently when they wish to expand production. The automotive industry represented around one-tenth of global semiconductor demand in 2020 and only 3% of total revenues of the Taiwan Semiconductor Manufacturing Company in 2019 — the world's largest foundry and top supplier of chips to the automotive industry. By comparison, communication and computers, the two largest semiconductor markets by end-use, jointly account for two-thirds of total demand. In this context, given that it is likely to take semiconductor makers at least six to nine months to realign production, the current supply shortage for motor vehicle manufacturers may continue well into 2022.

← 1. Motor vehicle production accounts for three-quarters of the overall output of the transport equipment industry at the world level.

← 2. Even though the share of electronic and electrical equipment inputs imported from Asia and Pacific is relatively small in the North American motor vehicle sector, US carmakers can be highly reliant on Asia for some critical parts and components. Many car producers reported that they had to reduce or cut production this year due to shortages of electronic chips. Part of these chips are produced in the United States, but the bulk of the imports is sourced from a small number of Asian countries.

Supply bottlenecks and labour shortages are checking the momentum of the recovery

The rapid recovery in global demand over the past year and the slower recovery of production capacity in some sectors have generated supply shortages. While new COVID-19 infections and deaths have been reduced substantially in many advanced economies, outbreaks are continuing to occur in some parts of the world, including many European economies in recent weeks, extending some supply constraints and creating new ones. Such constraints appear to be a factor in the loss of momentum visible in recent economic indicators. Shutdowns and other sanitary restrictions from the continued spread of the COVID-19 virus around the world this year have contributed to the persisting supply disruptions that are holding back the recovery (Figures 1.7 and 1.8), and are putting strong upward pressure on some prices. Survey indicators of supplier delivery times have risen to very high levels in many advanced economies (Figure 1.8, Panel A), particularly in Europe and North America, and inventory levels have declined substantially in many industries. Business surveys suggest that the majority of firms do not expect supply disruptions to be resolved before the latter half of 2022 (Figure 1.7, Panel A; de Bondt et al., 2021; IFO, 2021). COVID-19 outbreaks have led to closures of key ports, such as Shenzen and Ningbo in China, creating bottlenecks in global shipping and slowing container traffic (Figure 1.8, Panel B). The pandemic has also been also a factor responsible for closures of factories producing computer chips, which in turn has curbed the production of goods using those chips, especially cars (Box 1.1).

In labour markets, signs of shortages have also appeared (Figure 1.9), even though overall employment has yet to recover fully from the pandemic. In part, this may reflect changes in the location of activities, changes in the skill mix required in the context of the pandemic (notably arising from the sudden shift in consumption from services to goods in many economies and the greater importance of on-line sales), or changes in matching efficiency. This can also be seen in the unfavourable shift in the relationship between vacancies and unemployment – the Beveridge curve – in some economies, such as the United States and Australia. In contrast, in economies like Germany and France, where job retention programmes were relatively broad and employer-employee matches have been preserved, Beveridge curves appear to have been little changed in the aftermath of the pandemic.

The pandemic also led some people to withdraw from the labour force, sometimes by choosing to retire early. Among OECD countries, the largest falls in labour force participation rates have been in Latin America, with the United States, Turkey and Israel also having large declines (see Figure 1.20), By contrast, Japan and a number of European economies had higher labour force participation rates in the latest quarter than they did just before the pandemic. Shortages have also emerged in sectors and countries normally reliant on sizeable cross-border inflows into the labour force. Permanent migration to OECD economies declined by around 30% in 2020, and temporary labour migration also fell sharply (OECD, 2021a).1 The need to re-hire workers as sectors like hospitality and travel progressively reopened has thus put upward pressure on wages (Figure 1.10), especially in countries like the United States where there was less preservation of firm-worker matches. Despite an increase in near-term inflation expectations, however, expected wage growth remains similar to pre-pandemic norms (Figure 1.11).

Figure 1.7. Supply chain disruptions are affecting many firms and are expected to persist

Source: European Commission; Duke University, FRB Richmond and FRB Atlanta CFO Survey Q3 2021; and OECD calculations.

 StatLink https://stat.link/zo7s6d

Figure 1.8. Supply constraints have pushed up delivery times and slowed global trade

Note: Series in Panel A based on Manufacturing PMI and defined as 100 minus the proportion of PMI survey respondents reporting that delivery times have stayed the same or become faster.

Source: Markit; RWI/ISL Container Throughput Index; and OECD calculations.

 StatLink https://stat.link/0mhju7

There have been a number of other significant disruptions this year, many of which are related to climate change. Huge wildfires (notably in Siberia – possibly the largest in recorded history – California and Turkey), unprecedented heatwaves and droughts (e.g. in western North America), extreme cold weather events and destructive floods (e.g. in Germany, Belgium and western Canada) have collectively caused thousands of fatalities and major destruction of property and economic disruption. Hurricane Ida in late August and early September 2021 was one the costliest storms in US history and significantly restricted the output and transportation of oil and gas for many weeks, adding to upward pressure on global energy prices (Box 1.2). Such storms, along with other weather-related disasters, have become more frequent and severe as sea and air temperatures rise (World Meteorological Organisation, 2021; IPCC 2021). The economic impact has increased correspondingly, especially on agriculture (FAO, 2021), with the low-income countries being the worst affected.

Figure 1.9. Labour shortages have emerged at an early stage of the recovery

Note: Data in Panel A are a three-month moving average.

Source: National Federation of Small Business; European Commission; and OECD calculations.

 StatLink https://stat.link/axpcre

Figure 1.10. Aggregate wage pressures remain moderate despite large rises in some sectors

Note: The wage measures in Panel A are: the median change in hourly wages of individuals observed 12 months apart in the United States; contractual cash earnings in establishments with five or more employees in Japan; and negotiated wages excluding one-off payments in Germany.

Source: Federal Reserve Bank of Atlanta; Bureau of Labor Statistics; Destatis; Ministry of Health, Labour and Welfare, Japan; and OECD calculations.

 StatLink https://stat.link/v821hj

Figure 1.11. Inflation expectations have moved up, but expected wage growth is little changed

Note: In Panel B, ‘Selected other OECD countries’ corresponds to a weighted average of other OECD countries for which comparable wage expectations data are available: Australia, Canada, New Zealand, Norway, Sweden and the United Kingdom. Countries are weighted by employment levels. The value for 2021Q4 is based on the October value for the United States, and is available only for New Zealand, Norway and the United Kingdom for the aggregate.

Source: Refinitiv; Chartered Institute of Personnel and Development; Bank of Canada; and OECD calculations.

 StatLink https://stat.link/tlis8n

COVID-19 related and other supply disruptions have helped to push up commodity prices over the past year during a time when demand for these commodities, albeit rising rapidly, was still lower in volume terms than before the pandemic. Oil prices more than doubled between June 2020 and late November 2021, a period during which global production was consistently below 2019 levels. Similar patterns were seen for coal, some metals, and a range of agricultural and industrial commodities. Gas prices and electricity costs have also soared. The spike in energy prices in recent months, resulting from a complex mix of causes (Box 1.2), is directly felt by households, helping to explain why their expectations of near-term inflation have risen strongly in many countries.

Box 1.2. The recent rise in energy prices – drivers and implications

Prices of all major energy commodities have risen sharply since mid-2020

The recovery of energy prices over the past 18 months – after some future prices for oil and natural gas had turned negative in the first wave of the pandemic – has been spectacular. Spot prices for Brent crude have doubled since end-June 2020, while the increases for coal and natural gas have been much larger still, by around 8 times and 18 times respectively at their peaks in early October, with some retracement since then (Figure 1.12, Panel A). Most of the increase has occurred since the early part of this year.

The common factors in the surge in the prices of different energy commodities are restricted supply and the resurgence of demand as economies progressively reopened (Fernández Alvarez and Molnar, 2021). Beyond those basic shared features, however, there have been a range of specificities for the different commodities and particular markets.

The number of active oil rigs globally has risen by about half since falling to a record low in mid-2020, following the sharp decline in oil prices, but it remains well below the levels prevailing just before the onset of the pandemic. The OPEC+ agreement in April 2020 initially reduced global oil output by around 10%, with a partial unwinding of the production cuts thereafter. This was accelerated via subsequent agreements in 2021 as oil demand recovered. Global supply this year has also been restrained by such events as the Texas freeze in February and Hurricane Ida in August. More recently, oil prices have surged to their highest level since late 2014, in part because of additional demand for power generation.

Figure 1.12. Fossil fuel prices have surged, pushing up electricity generation costs

Note: In Panel A, monthly averages from daily prices (November is based on data up to 24 November). Coal price is Australian steam coal average FOB Newcastle (spot price reported by HWWI); gas is Title Transfer Facility Netherlands daily futures price; oil is Brent crude spot price. Panel B shows the contributions from changes in commodity prices and changes in carbon permit prices to changes in the cost of electricity generation using gas and coal respectively in the European Union between June 2020 and October 2021.

Source: Refinitiv; US Energy Information Administration (EIA); and OECD calculations.

 StatLink https://stat.link/5nfbj4

In response to the drop in global demand and low prices, global coal production fell by 5% in 2020, its largest decline since the Second World War. The recovery in supply as demand rebounded was held back by a number of factors including pandemic-control measures (China), floods (Australia and Indonesia) and rail disruption (South Africa).

Significant imbalances between demand and supply have emerged in gas markets, and put significant pressure on international gas prices, even though global demand has returned only recently to pre-pandemic levels. Chinese demand for liquefied natural gas (LNG) has remained unusually high since the winter of 2020, causing some gas that would otherwise have gone to Europe to be diverted. Low water levels and hydroelectric power generation in South America in early 2021 further boosted international LNG demand. The European winter was also colder than normal, spurring greater demand for gas and causing an unusually large draw-down of stocks that have not been restored fully this year. Restocking of reserves in Russia has been a factor in limiting supplies to Europe as demand has turned up this year. European demand for gas in the summer of 2021 was also pushed up by low production of wind power, with wind speeds across the continent unusually low.

In jurisdictions where carbon permit prices have also increased (especially Europe, including the United Kingdom, which now has its own permit system), this has contributed modestly to rising energy prices. The EU carbon permit price roughly tripled from its level just before the pandemic struck to its level in mid-October 2021, though this accounts for only a small part of the total increase in the cost of electricity generation using gas or coal (Figure 1.12, Panel B; see also Pacce et al., 2021).

Rising energy costs have hurt households’ purchasing power and fed the rise in inflation expectations

The direct impact of higher prices for oil, natural gas and coal on household expenditures has so far been limited given that spending on electricity and fuels is in the range of 5-10% in most OECD economies and that not all of the increases in spot prices have yet fed through into retail prices. Transaction prices, especially for natural gas, are much less volatile than the spot price, as a large share of gas is shipped on the basis of long-term contracts. In addition, a large share of retail prices reflects costs not related to commodity prices, such as the cost of labour in transportation and indirect taxes. Thus, for example, while the price of crude oil in the United States more than doubled over the year to mid-October 2021, retail gasoline prices increased by only about 50% during the same period. Finally, in some countries, governments have intervened to hold down the prices faced by households. About one third of OECD countries have taken action to cushion the rise in electricity prices, through a mixture of price caps, tax reductions and support for lower-income households (Boone and Elgouacem, 2021).

Even if, overall, the impact of the increase in energy prices on household budgets so far has been limited, that is not necessarily the case for all households. Lower-income households tend to spend a higher share of their income on energy (Figure 1.24), and they are less able to cushion the effect of a jump in energy costs by drawing on savings or reducing other expenditures, as they also spend proportionately more on other necessities.

The contribution to consumer price inflation from energy prices should peak soon

If the price increases for oil and natural gas since mid-2020 are not unwound, further rises in prices for electricity, gas, heating oil and motor fuels can be expected as higher commodity prices become reflected in long-term contracts and regulated price ceilings are raised. In addition, the rise in energy prices has raised costs for businesses already, particularly for industries such as chemicals and metals, and may continue to be passed on into output prices. Even so, however, without further rises in commodity prices, the biggest hit to household budgets will have occurred by the end of the winter of 2021. Further near-term fuel price rises are possible, given relatively low inventory levels, but a likelier outcome is that the supply constraints that have held output below pre-pandemic levels are unwound. In time, demand will also adjust to pricing anomalies that have emerged; for example, at its peak in October 2021 the oil-equivalent price of natural gas in Europe was around USD 200 per barrel, incentivising the use of other fuels. Apart from substitution between gas, oil and coal, the price spikes that have occurred could also help to spur faster development of alternatives to fossil fuels.

Headline inflation and, to a lesser extent, core consumer price inflation have risen markedly in most countries over the past year (Figure 1.13). Among large advanced economies, this is particularly the case for the United States, as well as, to a lesser extent, the United Kingdom, Germany and Canada. Commodity prices have risen particularly sharply, as have prices in many durable goods sectors, especially in the United States, where the switch in consumer demand from services to durable goods was particularly marked (see below). There was much less of a surge in durables demand in the euro area, for example, and price spikes for durable goods were accordingly much more muted. Core inflation (excluding food and energy) in the median advanced economy remains close to or below central bank targets, at 2.2% in October. Base effects are also affecting fluctuations in annual inflation. Some prices fell in most countries in the early phase of the pandemic, creating a low base for year-on-year increases in 2021. In many economies, inflation when smoothed over a two-year period, to control for the pandemic and the rebound, has been similar to the preceding two years, although the United States is a notable exception. Finally, large pandemic-induced changes in consumption patterns and relative prices have created some challenges to the measurement of inflation (Box 1.3).

Financial conditions have generally remained very favourable worldwide in recent months, and have tended to evolve in a similar manner in large advanced economies. In contrast, developments have been more diverse in emerging-market economies. The signal by the US Federal Reserve in September that the tapering of asset purchases could begin imminently prompted repricing, with US nominal bond yields rising for some weeks. Still, 10-year government bond yields in late-November were little changed from May in the United States,2 the euro area, Japan and the United Kingdom, which suggests that longer-term inflation expectations have remained anchored despite rising inflation. In contrast, 10-year yields were over 1 percentage point higher than in May in some large emerging-market economies, including Brazil, Turkey and Russia (Figure 1.14). In the euro area, spreads relative to Germany have remained broadly stable. In most economies, both advanced and emerging-market, equity prices generally remain higher than earlier this year. Most large advanced economies’ currencies have depreciated against the dollar since May, while the picture is mixed for the major emerging-market economies.

Figure 1.13. Headline inflation has increased but is mainly concentrated in goods

Source: OECD Consumer Price database; Statistics Bureau, Japan; National Statistical Office, India; and OECD calculations.

 StatLink https://stat.link/bjao02

Figure 1.14. Financial market conditions generally remain favourable

Note: 'Latest' refers to the change between May 2021 and the latest available data up to 24 November. 'Max / min' refers to the maximum change since May 2021. Based on a 10-day average of daily observations.

Source: Refinitiv; and OECD calculations.

 StatLink https://stat.link/8id3oq

Box 1.3. Inflation measurement during the pandemic and beyond

The COVID-19 crisis has complicated inflation measurement. One difficulty is that the implementation and subsequent lifting of lockdowns and social-distancing measures have significantly altered consumption patterns in most economies. The Consumer Price Index (CPI) is calculated as a weighted average of the prices of different goods and services, with weights based on past expenditure shares. If consumption patterns undergo sudden and large changes, those weights become unrepresentative of the latest expenditure shares, and gaps can emerge between inflation computed with official versus “real-time” weights. The likelihood of sizeable gaps rises when there is a high dispersion of price developments for different categories of goods and services.

An illustrative estimate of the potential magnitude of those measurement gaps for the United States, Japan and the euro area can be obtained by comparing a counterfactual (or “real-time”) CPI generated by retrospectively applying weights known at a future point in time (OECD, 2021b) with a CPI generated using official weights.1 The exercise is done for 2019 (the pre-pandemic year), 2020 (when lockdowns were generally the strictest) and 2021 (when some return towards more normal patterns has been underway in most economies). In 2020, the weights of those expenditure categories requiring mobility or personal contact fell significantly, while the share of food and housing in households’ budgets typically increased. Real-time weights for 2021 are not yet known, but different assumptions can be made about how fast and to what extent consumption patterns return to their pre-pandemic norms. One possible scenario is that those patterns fully go “back to normal”, with pre-pandemic weights holding again (Table 1.2). This is a strong simplifying assumption, since the composition of consumers’ expenditure in 2021 remains affected by the pandemic, and a full return to pre-pandemic patterns may not occur (Hodbod et al., 2021).

Table 1.2. Assumptions on the base years for weights
	Scenarios	2019	2020	2021
United States CPI	COICOP-adjusted official weights	2018 weights (released in early 2019)	2019 weights (released in early 2020)	2020 weights (released in early 2021)
	counterfactual (or “real-time”)	2019 weights (released in early 2020)	2020 weights (released in early 2021)	2019 weights (“back-to-normal”)
Japan CPI	COICOP-adjusted official weights	2015 weights (released in mid-2016)	2015 weights (released in mid-2016)	2020 weights (released in mid-2021)
	counterfactual (or “real-time”)		2020 weights (released in mid-2021)	2015 weights (“back-to-normal”)
Euro area HICP	official weights	2018 weights (released in early 2019)	2019 weights (released in early 2020)	2020 weights (released in early 2021)
	counterfactual (or “real-time”)	2019 weights (released in early 2020)	2020 weights (released in early 2021)	2019 weights (”back-to-normal”)
Note: COICOP (Classification of Individual Consumption According to Purpose) is the international reference classification of household expenditure. To improve comparability across economies, COICOP-adjusted weights and indices are used (computed by the OECD with the cooperation of national statistics offices), rather than the country-specific classifications used by the United States and Japan (the euro area HICP already uses COICOP). For the United States, the 2020 weights are only price-updated weights, and thus do not fully capture the impact of the pandemic on consumption patterns.
Source: OECD Main Economic Indicators database.

Following the methodology set out in OECD (2018), the contributions of twelve 2-digit COICOP expenditure categories to the overall annual inflation rate under different weights have been computed (Figure 1.15).

• In the United States, “real-time” inflation has been above the inflation rate computed using COICOP-adjusted official weights since spring 2020. Inflation in transport services was below average in 2020, so the higher official weight (based on 2019 data rather than the decline in transport services spending in 2020) led to a stronger downward impact on aggregate inflation. Conversely, when transport prices accelerated in spring 2021, the higher “back-to-normal” weight has generated a larger contribution to aggregate inflation. Thus there has been a consistently positive contribution of transportation to the difference between counterfactual and official inflation.2

• In Japan, “real-time” inflation has persistently been above the inflation rate computed using COICOP-adjusted official weights, mainly due to the contributions of “Restaurants and hotels” from 2020 and “Communication” from April 2021. The weight of “Restaurants and hotels” fell in 2020 while prices fell in 2020 and recovered in spring 2021. In the case of “Communication”, a significant fall in mobile phone charges in 2021 (largely unrelated to the pandemic) has exerted a stronger downward impact on aggregate inflation when the higher official weight (referring to 2020) is used.3

Figure 1.15. Differences between annual inflation rates computed with counterfactual and official weights

Note: The bars display, for each 2-digit COICOP category, the difference between the contributions to annual all-items inflation obtained using counterfactual and official weights. The category “Other goods and services” consists of “housing, water, electricity, gas and other fuels”, “furnishing, household equipment and routine maintenance”, “health”, “education”, and “miscellaneous goods and services”. Triangles represent the difference between all-items inflation rates derived by aggregating the contributions of all categories.

Source: OECD Main Economic Indicators database; and OECD calculations.

 StatLink https://stat.link/j07p61

• In the euro area, the “real-time” inflation rate has often been higher than the inflation rate computed with official weights, though with some fluctuations. “Transport” and “Restaurants and hotels” both have lower weights in 2020. However, unlike Japan, price increases in “Restaurants and hotels” in 2020 were generally above average, resulting in a negative contribution to the difference between this counterfactual and inflation computed using official weights.

Overall, this illustrative analysis suggests that sharp changes in consumer behaviour, captured by large shifts in expenditure weights, could currently be leading to some underestimation of annual inflation. However, as in OECD (2021b) and ECB (2020), the impact of different expenditure weights tends to be modest: in the present analysis the upper limit of this impact is only around 0.15 percentage points in the United States, 0.35 percentage points in Japan, and 0.2 percentage points in the euro area.4 The higher estimates for the United States reported in Cavallo (2020) largely stem from the use of weights that are updated monthly, based on credit and debit card transactions. This captures the exceptional and temporary shifts in consumption patterns at the height of the pandemic, but is less comprehensive than households’ total consumption expenditure.

In order to assess the robustness of the above conclusions, a related analysis has been carried out at more disaggregate levels (3-digit and 4-digit COICOP expenditure categories) for the euro area. (COICOP data at this level are not available for the United States and Japan). The difference between annual inflation rates computed with “real-time” and official weights becomes considerably more volatile, partly as a consequence of the marked seasonality in the prices of some smaller items with large changes in weights, such as package holidays and accommodation services. These factors are aggregated when two-digit weights are used, reducing the overall volatility. However, when averaged over each year, the underestimation of inflation is only marginally larger (0.17 percentage points at most) than that computed at the 2-digit level.

← 1. This index may differ from the official CPI due to a number of methodological differences relating to chain linking and because the consumption classification used in this box (COICOP) differs from the classification used in the CPIs of the United States and Japan.

← 2. In other words, the decrease in the “real-time” weight in 2020 weakened the impact of the price fall, while the recovery in the “real-time” weight in 2021 amplified the impact of the price rise, resulting in a consistently positive contribution.

← 3. It should be noted that this higher weight mainly reflects the greater usage of smartphones since 2015, rather than pandemic effects.

← 4. The extent of the underestimation can vary depending on which “real-time” weights for 2021 are assumed. The estimates presented in this box assume a fully “back to normal” situation. If the weights only moved half the way towards 2019 consumption patterns, the size of the underestimation would also be roughly halved.

The evolution of the pandemic remains uncertain

Significant uncertainty remains about the evolution of the pandemic. On the upside, faster global progress in deploying effective vaccines would boost confidence and spending by consumers and companies and encourage an unwinding of the additional household savings accumulated since the outset of the pandemic. This would improve growth prospects in 2022 and 2023, with global output returning to the path expected prior to the pandemic and unemployment falling more sharply in the typical economy. At the same time, stronger demand could raise financial market expectations of an earlier start to interest rate increases in the advanced economies, which could create particular difficulties for some emerging-market economies. A major downside risk is that the speed of vaccine deployment, and the effectiveness of existing vaccines, will be insufficient to stop the transmission of COVID-19 variants of concern, resulting in a need for new or modified vaccines or repeated campaigns to provide booster doses. In such circumstances, stricter containment measures might need to be re-employed (as is currently occurring in a number of European countries), lower-income economies could continue to lag behind in vaccine deployment, and output and per capita incomes would remain weaker than the pre-crisis path for an extended period. New mobility restrictions and port closures could hamper global trade, as persisting shutdowns in key economies reduce the availability of supplies along supply chains and lengthen supplier delivery times. Such further supply disruption might also create additional upward pressure on some prices.

Inflation could continue to surprise on the upside

Headline consumer price inflation is projected to peak in the majority of advanced and emerging-market economies by the first quarter of 2022, before moderating gradually (Figure 1.21). Large increases in commodity prices, including recent jumps in energy costs, supply shortages, and higher transportation costs are key factors that have pushed up prices around the world. In the major economies, higher shipping costs and commodity prices are estimated to be adding over 1½ percentage points to consumer price inflation in the latter part of 2021, accounting for around three-quarters of the jump in inflation since the fourth quarter of 2020 (OECD, 2021c). The impact of these input price rises on consumer price inflation is expected to fade gradually through 2022-23, with key bottlenecks easing as capacity expands and the rebound in consumer demand growth moderates. Broader cost pressures, particularly in labour markets, are projected to remain moderate but pick up slowly as the recovery proceeds, with unit labour costs in the OECD economies rising by around 1¼ per cent in 2022 and 1¾ per cent in 2023.

Figure 1.21. Inflation is projected to peak by early 2022 in most countries but remain above pre-pandemic levels

Note: The harmonised consumer price index for the euro area; the personal consumption expenditure deflator for the United States; and the overall consumer price index for the remaining countries shown separately.

1. Advanced economies include OECD countries except Chile, Colombia, Costa Rica, Mexico and Turkey.

Source: OECD Economic Outlook 110 database; and OECD calculations.

 StatLink https://stat.link/4pv25u

• For 2022 as a whole, the consumer price inflation across the OECD is projected to rise to 4¼ per cent, and to 3½ per cent in the major advanced economies. As the effects from past input price rises drop out, and supply constraints wane, inflation is projected to moderate to around 2% in the typical advanced economy by 2023. Inflation is projected to remain above 2% in the United States, the United Kingdom and Canada, prompting some gradual tightening of monetary policy. In the euro area, inflation is projected to settle between 1¾-2 per cent by 2023.

The risks around the baseline scenario are high. It is possible that the resolution of supply constraints will permit a faster-than-projected unwinding of some of the sharp price increases seen over the last year, resulting in a faster decline in inflation. Such an outcome would be benign for the global economy, although it would heighten the challenge of achieving monetary policy objectives in countries that already have underlying inflation below central bank targets.

The main risk, however, is that inflation continues to surprise on the upside, forcing the major central banks to tighten monetary policy earlier and to a greater extent than projected. Such an outcome could stem from a number of possible factors, including prolonged supply disruptions, an upward shift in inflation expectations, labour market pressures, or if prices for a wider range of goods and services start to rise substantially.

• Existing supply disruptions could continue to be more severe, prolonged and widespread than expected, or new sources of bottlenecks could emerge. The most likely cause of such a continued impairment of supply chains is a worsening of COVID-19 dynamics, whether globally or in specific countries that are important to global value chains, particularly in Asia. A renewed worsening of the pandemic could lead to a repeat of the shift in demand from services to goods seen in many countries in the first phase (Figure 1.23), creating another squeeze in goods markets with restricted supply and surging demand, and giving further impetus to inflation. Continued disruptions to energy supply that help to keep energy prices high for longer would also heighten inflation pressures (see below).

• Inflation expectations may become entrenched at levels above central bank targets. Household inflation expectations in many countries have moved up as realised inflation has increased, especially over a one-year horizon. Business expectations of inflation have also generally increased, though by less. Measures of expectations derived from asset prices have tended to show only a modest increase, especially over longer horizons (5-10 years). Household and corporate wage expectations also generally remain moderate, though marginally higher than before the pandemic (Figure 1.11).

• Another potential consequence of continued upside inflation surprises is that the acceleration in price increases becomes broad-based. Past experience in OECD countries suggests that increases in inflation associated with sharp rises in a relatively small number of items tend to be short-lived, while the longer-lasting inflationary periods are characterised by a more-uniform distribution of price increases across components of the index. One common method for eliminating the influence of large but narrowly focussed price changes is to calculate trimmed means, but central banks employ varying methodologies to calculate such measures. Estimates of trimmed mean inflation across a number of OECD countries using a consistent methodology suggest that there is some broad-based inflationary pressure in the United States and the United Kingdom, but less in the euro area or Japan (Box 1.4).

• A particular source of risk for a broadening of inflationary pressure is housing. Housing costs, primarily rents, account for a large share of household expenditures. Historically, increases in housing costs have tended to be very stable, and so far little of the upward pressure on headline inflation has come from this source, but the large weight of housing in price indices means that there is scope for inflation to move higher if rents begin to rise more quickly.5 Housing, together with food and energy, which have been major sources of upward pressure on consumer prices over the past year, also comprises a larger share of expenditures for lower-income households (Figure 1.24). Inflationary pressure in these components is likely to be keenly felt by many households, potentially contributing to higher inflation expectations and higher wage demands.6

• Higher inflation expectations, tighter labour markets, or skill shortages could all trigger stronger wage pressures than projected currently, leading to further price increases or squeezed profit margins. These types of linkages were seen in the 1970s, with upward pressures on inflation from a number of different factors, including the oil price shock in 1973-74, leading to persistently higher wage and price inflation for several years. Changes in labour market institutions since the 1970s have, however, reduced this risk in many countries, though not eliminated it, with a decline in the coverage of collective bargaining agreements, the removal of many automatic wage indexation mechanisms and a reduction in employees’ bargaining power due to lower union membership.7 In the medium term, there are also risks that some of the structural forces helping to place downward pressure on inflation in the past two decades may reverse. In particular, a reconstitution of global supply chains by moving some activities to closer but higher-cost locations would unwind some of the effects on inflation from the globalisation of economic activities since the mid-1990s (Koske et al., 2008; Andrews et al., 2018). Competitive pressures in product markets might also weaken, including from abroad, in the absence of further reforms to allow market exit and entry in the aftermath of the pandemic.

• International mobility restrictions and the large decline in cross-border migration since the onset of the pandemic could also have long-lasting consequences on labour market pressures as sizeable net migration inflows have been an important source of labour force growth over a long period in many countries. Migrants tend to be concentrated in certain sectors, with foreign-born labour accounting for over 10 per cent of the manufacturing labour force in many OECD economies, as well as a high number of seasonal workers in agriculture. A reduction in migration flows can thus increase wage pressures at the sectoral level. For developing economies, restrictions on cross-border mobility also tend to reduce remittances from abroad. These important flows could decline by 14% in 2021 relative to pre-COVID levels (World Bank 2021).

Box 1.4. Estimating underlying inflation pressures using trimmed means

A key challenge for policymakers is to be able to separate persistent changes in price pressures from changes that are large but ultimately short-lived. Headline inflation is the aggregation of the changes in hundreds of prices of products that are part of household consumption expenditure. Underlying inflation is based on a subset of these changes that are thought to contain information about potential future inflation developments. A good measure of underlying inflation should be less volatile than headline inflation, unbiased, and capture the trend, so that headline will tend to adjust towards underlying inflation (Roberts, 2005).

One widely used estimate of underlying inflation is the core inflation rate, an exclusion-based measure, which removes energy and food-related items in the consumer basket. This works well when temporary fluctuations in commodity prices are the main source of volatility in inflation, but may understate price pressures when there are large fluctuations in a broader range of goods and services.

Trimmed-mean inflation is an alternative measure of underlying inflation pressures. This approach is purely statistical and aims to remove the most extreme price changes on both sides of the distribution of price changes each month, and then to compute mean inflation from the remaining items. While core inflation always removes the same products, those considered to have inherently volatile prices, the trimming eliminates the most extreme price changes in every period, which may involve taking out different items each month.

One practical difficulty is determining the optimal degree of trimming: removing too few items could leave excessive volatility, while removing too many could reduce the usefulness of the measure and push trimmed-mean inflation away from the (unobserved) trend in inflation (Dolmas and Koenig, 2019). Trimmed-mean measures are widely used by the major central banks, but there is no consensus on the optimal degree of trimming. For example, in the United States the Federal Reserve Bank of Cleveland (2021) trimmed-mean CPI inflation estimate removes 8% of the weight on each side of the distribution, and the Federal Reserve Bank of Dallas (2021) trimmed-mean personal consumption expenditure inflation measure is asymmetric, with 24% of the weight from the lower tail and 31% of the weight in the upper tail being trimmed. The Bank of Japan (2021) and Bank of Canada (2021) have been using a symmetric trimming of 10% and 20% at the top and bottom of the distribution, respectively. The European Central Bank (ECB) does not systematically publish trimmed-mean inflation numbers but has been recently using a symmetric 7.5% trimming (ECB, 2021).

The measures below use a symmetric trimming of 10% at the top and bottom of the distribution to facilitate comparisons across countries, and to reduce the difference with a 36-month average inflation rate (which is used as a reference estimate for trend inflation).

Both core and trimmed-mean inflation are less volatile than headline inflation (Figure 1.22). They usually move in a similar fashion when the most volatile items are food and energy products. During the first phase of the pandemic, headline inflation declined substantially, but a comparison between core and trimmed-mean inflation reveals different drivers across countries. More recently, trimmed-mean inflation has been weaker than headline inflation (except for Japan) but on an upward trend, pointing to a gradual broadening of underlying inflation pressures.

• In the United States, trimmed-mean inflation barely moved in 2020 while core followed headline inflation. The observed deceleration in headline inflation in 2020 thus came from a limited number of items not related to food and energy, most likely related to sanitary restrictions on some services. More recently, annual trimmed-mean inflation has increased quite sharply to 4.1% in October 2021. Core inflation also moved up to 4.1% in October after having stabilised at around 3.6% in the previous four months. These developments suggest that the rise in prices has become more broad-based in recent months.

Figure 1.22. Trimmed-mean estimates of inflation pressures are now rising

Year-on-year percentage changes

Note: Data are for the personal consumption expenditures deflator for the United States; consumer price inflation for Japan; and harmonised consumer price inflation for the euro area and the United Kingdom. Trimmed-mean inflation trims 10% in terms of weights at the top and bottom of the distribution of the year-on-year growth of prices. Core inflation excludes energy and food-related products.

Source: OECD Economic Outlook 110 database; Bureau of Economic Analysis; Japan Statistics Bureau; Eurostat; Office for National Statistics; and OECD calculations.

 StatLink https://stat.link/xyjak8

• In Japan, trimmed-mean inflation has usually been above core inflation since 2020, in a period of low headline inflation. This suggests that a small number of components with a large negative inflation rate, outside of food and energy, are holding down headline inflation. Trimmed-mean inflation, which excludes those items, has risen to 1.2% in October.

• In the euro area, during the pandemic core and trimmed-mean inflation were similar, as the same components were removed in both measures. More recently, headline inflation has risen by more than both underlying inflation measures, highlighting the role of higher food and energy costs. However, underlying inflation has also risen, with trimmed-mean inflation reaching 2.4% in October.

• In the United Kingdom, in 2020, as the pandemic hit the economy, headline inflation fell while core inflation remained more stable, the trimmed-mean measure being in the middle. This suggests that some domestic sectors as well as food and energy products had extremely low inflation rates. More recently, trimmed-mean inflation increased to 3% in October, pointing to rising underlying inflation pressures.

Figure 1.23. The pandemic triggered a shift in consumption from services to goods, especially in the United States

Real household consumption expenditures

Note: In Panel B, the data are an average of euro area member countries weighted by nominal consumption expenditures. Non-durable goods incorporate semi-durable goods. Services and non-durable goods are available only for Austria, Estonia, Finland, France, Germany, Ireland, Italy, Latvia, Luxembourg, the Netherlands and Portugal. 2021Q3 data are available only for France, Germany, the Netherlands and Spain.

Source: Bureau of Economic Analysis; Eurostat; and OECD calculations.

 StatLink https://stat.link/17vxw3

Risks remain that the combination of supply factors that have contributed to the gradual tightening of energy markets could persist and push up global energy prices further over the projection period (Box 1.2). A rise in energy prices reallocates income between producers and consumers. Higher energy prices raise production costs, leading to some energy-intensive capacity being scrapped, and push up consumer prices in all economies, but export revenues will rise, and so may investment, in energy-producing countries. The net impact of rising energy prices on global activity is expected to be negative, however, as the propensity to consume of energy importers is typically higher than that of energy producers.

Simulations using the NiGEM macro-model illustrate the implications of a further 30% rise in world oil, gas and coal prices, starting in the first quarter of 2022. The shock is assumed to be sustained for three years, before fading gradually thereafter. In the first two years of the shock, consumer price inflation is pushed up by around 0.9 percentage points per annum on average in the OECD economies, and by 1¼ percentage points per annum elsewhere (Figure 1.25). GDP growth declines in the OECD economies by a little over 0.3 percentage points per annum in the first two years, with smaller net declines in the non-OECD economies reflecting a balance between stronger growth in the energy-producing economies and deeper declines in the major importing economies. Global carbon emissions are pushed down initially due to reduced use of carbon-intensive energy, but this effect fades in the absence of a sustained price change. Monetary policy reacts to the upturn in inflation with policy interest rates raised by around 1 percentage point on average in the first two years in the major advanced economies before returning towards baseline as inflationary pressures subside.8

Figure 1.24. Food, transport and housing account for a larger share of the budgets of lower-income households

Shares of household budgets accounted for by food, housing and transport, by income quintile

Note: Final consumption expenditures by uses. The OECD represents the unweighted average of shares of Australia (for 2017), Canada (2019), Czech Republic (2017), France (2016), Ireland (2016), Israel (2017), Mexico (2018), Netherlands (2017), New Zealand (2015), Slovenia (2018), Sweden (2015), United Kingdom (2017) and United States (2016).

Source: OECD Distributional Information on Household Income, Consumption and Saving database (experimental); and OECD calculations.

 StatLink https://stat.link/0wd3iv

If the simulation is undertaken with backward-looking (adaptive) rather than forward-looking expectations for companies, households and financial markets, there are more noticeable differences. The first-year effect on inflation is similar, but thereafter the impact is weaker in the OECD economies. A key factor behind this is the larger negative effects on output, with GDP growth in the OECD economies declining by around 0.6 percentage points per annum on average in the first two years. This stems from a larger hit to investment and a reduction in energy use, with firms and consumers unaware that higher energy prices will ultimately not persist.

Figure 1.25. A further short-term rise in energy prices would hit growth and add to inflation

Note: Non-OECD oil denotes non-OECD oil-exporting economies and Non-OECD others denotes all remaining non-OECD economies. Forward (adaptive) indicates that producers and consumers have forward-looking (backward-looking) expectations. Simulation of a 30% rise in global oil, gas and coal prices sustained for three years and fading thereafter.

Source: OECD calculations using the NiGEM macroeconomic model.

 StatLink https://stat.link/hr0kgo

Risk repricing could expose vulnerabilities in the corporate sector and emerging-market economies

Financial markets have been resilient but vulnerabilities are mounting

Financial conditions have generally remained very favourable in both advanced and emerging-market economies. Reflecting ample policy support and robust investor risk appetite, bond issuance by sovereigns, non-financial corporations and financial firms has been very strong in 2021, and spending on mergers and acquisitions has soared.9 Spreads for lower-grade corporate bonds are still low, corporate debt levels have continued to rise (Figure 1.26, Panel A), and asset valuations appear very stretched in some markets, especially housing (Figure 1.27). These developments raise vulnerabilities to abrupt risk repricing in financial markets. The uncertain outcomes from the financial distress being experienced by China’s largest property developer could also trigger swings in risk appetite and slow global growth.

In contrast to the global financial crisis (GFC), current vulnerabilities seem to be more concentrated in the corporate sector. With the exception of China, household debt has been relatively stable over the last decade (Figure 1.26, Panel B) and household balance sheets are currently stronger than they were in 2007 (IMF, 2021a). More broadly, the GFC experience has played a role in limiting the amount of risk-taking in the household sector.10 Global corporate debt, on the other hand, has grown rapidly since the GFC, even as global corporate credit quality was declining (OECD, 2021d).

Figure 1.26. Debt levels have swollen, especially for companies

Note: GFC refers to the level of debt as of 2010 Q4. Debt comprises loans and debt securities.

Source: BIS Total Credit database.

 StatLink https://stat.link/2qn0x4

Figure 1.27. Housing valuations are stretched in many countries

House price-to-rent ratio, latest quarter available, long-run average=100

Source: OECD Analytical House Price database; and OECD calculations.

 StatLink https://stat.link/dhjpo0

Stress in the corporate sector is currently contained, but could rise

The significant public support provided to many companies since the start of the pandemic, favourable credit conditions, and the global recovery are still keeping corporate bankruptcies in check. In Europe, bankruptcies are well below their historical standards, especially in France and Germany, and a similar picture emerges in other advanced economies (Figure 1.28, Panel A). In the United States, the slowdown in corporate bankruptcies observed in 2020 has persisted through 2021. Even in sectors hit relatively hard, such as energy, consumer services, transportation or real estate, the fallout has been limited so far with a flow of bankruptcies close to pre-pandemic levels (Figure 1.28, Panel B). Major banks in the United States and Europe also posted record profits in the first half of this year, partly as a consequence of a reduction in loan loss provisions.11 Overall, the wave of bankruptcies feared at the start of the COVID-19 crisis has not yet materialised.

Figure 1.28. Corporate bankruptcies are still subdued

Note: For France and Germany, the bankruptcy index tracks the number of legal units that have started the procedure of being declared bankrupt by issuing a court declaration at any time during the reference quarter. The index covers firms in the industry, construction and services sectors (except public administration, defence, compulsory social security, activities of membership organisations, activities of households as employers and extra-territorial organisations and bodies). Data for the United States, Canada, Japan and Australia are taken from the OECD Timely Indicators of Entrepreneurship. Data for Canada include sole proprietorships. Data for medium and large US companies taken from S&P Capital IQ. The coverage is limited to relatively large US public and private companies with publicly-traded debt with assets greater than USD 2 million and private US companies with assets greater than USD 10 million. Hard-hit sectors cover energy, transportation, consumer services and real estate. The pre-COVID period ranges from September 2018 to March 2020.

Source: Eurostat; S&P Capital IQ; and OECD calculations.

 StatLink https://stat.link/6ewxc2

Corporate investment is picking up, even as debt keeps climbing. Since the start of the pandemic, the debt overhang has been a key source of concern, as high corporate debt typically tends to reduce investment in the aftermath of economic crises with negative implications for the recovery (Kalemli-Özcan et al., 2019; Demmou et al., 2021). The rapid debt build-up, however, has also been matched by a strong growth of liquid, short-term investments (cash or equivalent) held by companies (OECD, 2021d). This helps explain why rising debt has not prevented strong investment growth. Business fixed investment volumes surpassed their pre-recession level in the United States in the second quarter of 2021, a much quicker recovery than after some other recessions, with a steady rebound also in most other major advanced economies and in some large emerging-market economies.

Longstanding sources of risk remain, some of which have been intensified by the pandemic. The debt generated by the COVID-19 crisis could threaten the recovery in several ways. A large share of the global non-financial corporate debt stock is still rated either as “speculative” or as BBB, the lowest rating in the investment grade category. An unexpected growth slowdown or continued strong growth in their input costs could threaten firms’ repayment capacity and their ability to roll over debt. In a sample of mostly large public and private firms monitored by S&P Capital IQ, 5% of firms currently report an interest coverage ratio (ICR) below one, while 7% of firms report an ICR between one and two (Figure 1.29).12 A steep rise in funding costs or a sharp fall in earnings could therefore endanger a non-trivial share of the corporate sector.

Figure 1.29. Share of larger firms “at risk”

Note: Analysis based on the 28 000 firms covered by S&P for which the interest coverage ratio (ICR) is currently available. The “current” ICR is computed as the ratio between earnings before interest, taxes, depreciation and amortisation (EBITDA) and total interest expenses over the latest fiscal year (FY2020 or FY2021). The FY2019 ICR reports the same statistic for the same sample of firms. The sample covers medium and large public and private non-financial companies operating in OECD countries and major (non-OECD) emerging-market economies.

Source: S&P Capital IQ and OECD calculations.

 StatLink https://stat.link/6ba48g

Figure 1.30. Debt has increased substantially in many emerging-market economies

Note: Debt comprises loans and debt securities.

Source: OECD Economic Outlook 110 database; BIS credit statistics; and OECD calculations.

 StatLink https://stat.link/4nmo3x

Weak fiscal positions increase vulnerabilities in many emerging-market economies

A decline in COVID-19 infections, accompanied by a lifting of confinement measures, has strengthened mobility and enabled activity to return to pre-pandemic levels in many emerging-market economies. However, an often-slower pace of vaccination than in advanced economies implies that new waves of the virus might overwhelm health systems and bring back the need for mobility restrictions. Furthermore, a recent tightening of financial conditions, high debt, and deteriorating fiscal balances make many emerging-market economies vulnerable to potential swings in global risk sentiment.

The pandemic has resulted in a further rise in debt in emerging-market economies. Government debt has increased considerably, due to a collapse in tax revenues and elevated spending in response to the pandemic (Figure 1.30, Panel A). Emerging-market economy sovereigns issued USD 1.8 trillion of debt in the first half of 2021, 40% higher than the average amount issued in the first half of each year in 2017-19 (OECD, 2022). Emerging Asia accounted for half of that total, with China accounting for around one-half of the region’s issuance. The share of short-term obligations in non-investment grade debt issuance also rose, reducing average maturities, and foreign-currency-denominated debt issuance has declined, possibly suggesting more difficult access to foreign debt markets.13 The debt sustainability outlook in the low-income countries eligible for the G20 Debt Service Suspension Initiative has also remained complex, making it challenging for these countries to re-orient limited public spending towards health measures (IMF, 2021b). In some emerging-market economies, the rise in private sector debt has also been very substantial. Excessive leverage in the real estate sector in China, sizeable declines in commodity prices at the onset of the pandemic in Chile and large-scale credit guarantee measures and currency depreciation in Turkey have all added to non-financial corporate debt (Figure 1.30, Panel B).

Vulnerabilities in emerging-market economies are now skewed to fiscal imbalances rather than current account imbalances, in contrast to past episodes of shifts in investor sentiment and higher risk aversion, such as the “taper tantrum” in 2013 (Figure 1.31, Panels A and B). With a few exceptions, such as Chile and Colombia, the reliance on foreign investors in local-currency sovereign bond markets is now lower than in early 2013, which mitigates potential vulnerabilities to reversals in global risk sentiment (Figure 1.31, Panel C). However, large increases in public debt accompanied by limited demand by foreign investors could create doubts about debt sustainability and make it more challenging for businesses and households to raise new finance in domestic markets, hindering the economic recovery (Priftis and Zimic, 2020; Broner et al., 2021). Although the currency composition of public debt is often broadly comparable to early 2013, the higher share of foreign-currency debt in countries like Argentina, Bulgaria, Romania and Turkey (Figure 1.31, Panel D) may make these economies more vulnerable to abrupt fluctuations in exchange rates. Risks might also quickly intensify in some countries should a decline in private savings — enabled by the easing of confinement measures — and a pick-up in real investment lead to a deterioration in external balances, particularly if sizeable fiscal deficits persist.

Figure 1.31. Indicators of fiscal and external vulnerabilities in emerging-market economies

Note: Red bars in Panels A and B denote the average for 2020Q3-2021Q2. Green bars in Panels A and B denote the average of four quarters preceding 2013Q2, during which global financial markets volatility has increased. Countries in Panels A and B are sorted according to the average fiscal balance-to-GDP ratio during 2012Q2-2013Q1. Countries in Panels C and D are sorted according to the foreign ownership of local-currency government securities in 2013Q1.

Source: OECD Economic Outlook 110 database; IMF Sovereign Debt Investor Base for Emerging Markets database; and OECD calculations.

 StatLink https://stat.link/5bp71f

Capital inflows into emerging-market economies have recently moderated (De Crescenzio and Lepers, 2021) and government bond yields and spreads have started to pick up.14 Movements in local-currency government bond yields in emerging-market economies have recently displayed marked regional differences, with strong increases in Latin America and Europe, but not in Asia (Figure 1.32, Panel A). At the same time, foreign-currency sovereign bond yields have increased only mildly in all regions, suggesting that the repricing of sovereign risk in emerging-market economies primarily reflects currency risk rather than credit risk (Figure 1.32, Panel B).15 If depreciation risks were to materialise, they could delay the projected convergence of inflation to central banks’ target ranges, and potentially require further monetary policy tightening that would hamper the economic recovery.

Figure 1.32. Bond yields indicate a significant rise in currency risk in emerging-market economies

Note: Panel A shows the yield differential between 10-year, local-currency emerging-market economy and US government bonds. Panel B shows the JP Morgan EMBI global bond spread, a measure for sovereign risk spread of USD denominated emerging-market economy government bonds over US government bonds. The shaded area in both panels denotes March 2020. “EME – Europe” covers Bulgaria, Romania, Russia and Turkey. “EME – Asia” covers China, Indonesia, India, Malaysia, Philippines, Thailand and Vietnam. “EME – Latin America” covers Brazil, Chile, Colombia and Mexico. The “EME” aggregate covers all countries, additionally including South Africa. Aggregate spreads correspond to unweighted cross-country averages.

Source: Factset; Refinitiv; and OECD calculations.

 StatLink https://stat.link/0e5x9o

Risks of a sharp slowdown in China have risen

The property sector has been a strong source of growth for the Chinese economy for many years, but this has been accompanied by rising leverage of developers and, more recently, households. Regulatory efforts to curb banks’ property lending and strengthen the balance sheets of property developers were introduced last year. Recent events have highlighted the continuing risks in China’s real estate market, with the potential for large cross-sector and cross-border spillovers. The worsening financial soundness of some large Chinese real estate developers poses risks for financial intermediaries, including foreign ones. These would be amplified if the crisis in the sector were to worsen and lead to defaults and a substantial decline in property prices.

Any major default would be managed carefully by the government, potentially limiting the overall systemic risks to financial stability. Nonetheless, credit conditions could tighten and bond spreads increase, with significant risks to economic growth beyond those already incorporated in the projections. The construction sector and real estate services both represented between 7-7½ per cent of total value-added in 2020, with linkages stretching into many other parts of the economy (Rogoff and Yang, 2021). Not all of these activities will reflect new real estate development, as construction will also include non-housing activities and real estate services will include the management of existing occupied properties. A severe slowdown would, however, significantly impact other sectors, particularly ones providing materials for construction activity. Final demand would be hit directly, with real estate development representing around 25-30% of total fixed asset investment in 2019 and 2020. Regional and local governments could also face budgetary pressures if the revenue from sales of land rights were to decline substantially, potentially reducing spending on infrastructure investment. Greater uncertainty could also damp domestic demand.

Figure 1.33. A sharp slowdown in China would hit growth and trade around the world

Difference from baseline

Note: Simulated impact of a two-year decline of 2 percentage points per annum in domestic demand growth in China, a supply shock in China reflected in a rise of 10% in Chinese export prices, an increase of 50 basis points in global investment risk premia, and a decline of 10% in equity prices in all economies. The red bars show the contribution from the direct demand-driven slowdown in trade; the blue bars show the additional contribution from adding higher Chinese export prices; the green bars show the contribution from adding higher uncertainty; and the orange bars show the additional effects from lower equity prices. Commodity-exporting economies (CMY EXPs) include Argentina, Australia, Brazil, Chile, Indonesia, Russia, South Africa and other non-OECD oil-producing economies. East Asia includes Korea and the Dynamic Asian Economies.

Source: OECD calculations using NiGEM.

 StatLink https://stat.link/2cuv91

Illustrative simulations, using the NiGEM global macroeconomic model, highlight the different potential adverse effects that a slowdown in China might have:

• An unanticipated decline of 2 percentage points per annum in the growth rate of domestic demand in China for two years could directly lower global GDP growth by over 0.4 percentage points per annum, and global inflation by 0.2 percentage points per annum, with the impact on output in Japan, commodity-producing economies and other economies in East Asia being higher than elsewhere (Figure 1.33). Lower global commodity prices contribute to the decline in inflation, given the significant role that China has in many commodity markets, with energy prices down by around 5% and metals prices by around 10%.

• If the slowdown in China disrupted the production of key inputs needed in other countries – with the supply shock proxied by a 10% rise in Chinese export prices — the adverse effects on the world economy would intensify, with global GDP growth lowered by close to 0.7 percentage points per annum, Higher export prices from China would directly push up global inflation by around 0.4 percentage points per annum, illustrating the impact of supply bottlenecks, and inflation in the advanced economies by around ¼ percentage point per annum.

• If accompanied by a deterioration in global equity prices and heightened uncertainty, reflecting risk repricing in financial markets, global GDP growth could be lowered by 1 percentage point per annum on average in the first two years of the shock. These shocks would also be deflationary, with the combined impact of all the shocks considered reducing global inflation by around ¼ percentage point per annum in the first two years.

Downside risks to the trade outlook

While the major threat to the near-term trade outlook remains a resurgence of further COVID-19 outbreaks that result in new mobility restrictions and port closures, rising trade tensions and pressures around strategic sovereignty issues represent another risk. The near stalling of global trade in 2018-2019, due to geopolitical tensions, highlights the need for a stable and predictable environment that allows trade to thrive. Many countries and regions are currently pushing forward industrial policies for strategic technologies. A potential risk is that these policy measures could result in rising trade tensions or restrictions, especially because barriers to trade are still high compared to pre-2018.

Health policy priorities

A key multilateral priority is to ensure that all resources necessary are used to deploy effective vaccinations as quickly as possible throughout the world to save lives, preserve incomes and help overcome supply shortages in labour and product markets. The recovery will remain precarious in all countries until this is achieved. The majority of the world remains unvaccinated, with vaccination rates especially low in many low-income countries. Failure to ensure the global suppression of the virus raises the risks that further new, more-transmissible variants continue to appear, with containment measures having to be reintroduced.

Stronger international efforts are needed to provide low-income countries with the resources needed to vaccinate their populations for their own and global benefits. These include vaccine supply and assistance to help overcome domestic logistical hurdles to vaccine deployment, with an urgent strengthening of global ACT-A and COVAX mechanisms. Effective multilateral action is also required to share knowledge, medical and financial resources, and avoid harmful bans to trade. Such bans would be self-defeating given the strong cross-border linkages in supply chains for vaccines and healthcare products.

Monetary policy should remain supportive but gradually become less accommodative

Monetary policy in the major advanced economies remains highly accommodative, which is appropriate given that the recovery is still uneven and incomplete. Major central banks have communicated an intention to look through current increases in inflation and have so far left their key policy rates unchanged, though a number of smaller advanced economies have recently raised policy rates.16 In addition, the asset holdings of many central banks have increased further due to ongoing asset purchase programmes (Figure 1.34), though a number of central banks have started to slow down the pace of net purchases, or have even brought them to an end. The Federal Reserve has started to taper asset purchases given the state of the recovery, and the ECB has moderately reduced the pace of purchases under its Pandemic Emergency Purchase Programme (PEPP). The Reserve Bank of Australia ceased targeting the yield on the April 2024 government bond this November, whilst maintaining its government securities purchase programme. The Bank of Canada started to recalibrate and adjust the scale of its quantitative easing programme late last year and ended it in October 2021, with new government bonds now being purchased only to replace maturing bonds. In contrast, the Bank of Japan has extended the duration of its special funds-supplying operations until the end of March 2022 to facilitate financing in the private sector.

A key issue is the extent to which central banks will continue to maintain accommodative monetary policy and look through the current upturn in headline inflation. Tolerance of inflation overshooting is more likely if the factors pushing up inflation are seen as ones that will ultimately wane, underlying price developments are contained, and medium-term inflation expectations remain well anchored. Such judgements will become more challenging if current supply shocks and inflationary pressures turn out to be more persistent than currently expected. Current policy frameworks in the major central banks make it possible to accommodate some overshooting of inflation while preserving policy credibility:

• In August 2020, the Federal Reserve adopted a flexible average inflation targeting strategy, which tolerates inflation levels moderately over 2% for a period of time after a persistent period of low inflation.

• The new monetary policy strategy of the ECB, unveiled in July 2021, adopted a symmetric 2% inflation objective. Occasional periods during which inflation is moderately above target are compatible with this new strategy, and forward guidance has been strengthened.

• To enhance the credibility of its 2% inflation target, the Bank of Japan announced already in September 2016 an “inflation-overshooting commitment”. This can be regarded as a form of “make-up” strategy, in which past deviations from the inflation target are offset in the future.

Figure 1.34. Changes in asset holdings by major central banks since December 2019

Note: Data for the euro area are from end-September 2021. Private assets include corporate bonds, commercial paper, asset-backed securities and exchange traded funds. General government bonds are treasury bills and municipal, state and central government bonds.

Source: OECD Economic Outlook 110 database; Reserve Bank of Australia; Bank of Canada; Bank of England; European Central Bank; Bank of Israel; Bank of Japan; Reserve Bank of New Zealand; Sveriges Riksbank; US Federal Reserve; and OECD calculations.

 StatLink https://stat.link/25vtm4

To help anchor expectations and minimise the risk of an abrupt repricing in financial markets, clear communication is needed about the horizon and extent to which any such overshooting will be tolerated, together with guidance about the planned timing and sequencing of eventual moves towards policy normalisation. These steps should be sequential in the major advanced economies, with an initial stabilisation of central bank balance sheets (by only reinvesting proceeds from maturing assets) before policy rates are increased. In that respect, the Federal Reserve and a number of other central banks have announced a very clear distinction between balance sheet and interest rate policies and emphasised different criteria for their respective tightening.17 Such communication and guidance should help to minimise the risks involved in the process of policy normalisation, and enable normalisation to proceed gradually, as appropriate given still-high uncertainty.18 Asset purchases also come with specific costs and risks, which might justify tapering them before raising policy rates (Blinder et al., 2013), particularly if they amplify financial stability risks by driving down term premia.

An alternative approach, particularly at a time when inflation has risen unexpectedly quickly, might be to raise policy interest rates before ending net asset purchases of longer-dated securities. Such an approach would clearly signal an intent to reduce near-term inflationary pressures, while ensuring that the yield curve does not steepen excessively. The impact of a change in policy interest rates on the economy is also better understood than the effect of asset purchases, and might also be easier to reverse if needed. At the same time, this approach could send mixed signals, with little clarity about whether the intention was to reduce monetary policy accommodation, particularly as long-term interest rates are an important factor that helps to determine private sector spending in many economies. In effect, policy could be becoming more accommodative by expanding asset purchases and enhancing market liquidity, whilst seeking to damp inflationary pressures through higher short-term interest rates.19

However, even with a sequential approach starting with tapering of purchases, central banks could still be forced to increase policy rates earlier or faster than expected to counter persistently high inflation and signs that inflation expectations are drifting up substantially. If such a scenario gives rise to substantial repricing in financial markets, with potentially a rapid fall in asset prices, there could also be a need to implement additional net asset purchases temporarily to defuse market tensions and preserve the smooth transmission of monetary policy.

In the major advanced economies, monetary policy is projected to evolve gradually towards normalisation over the next two years in a sequential manner, though at a different pace across jurisdictions:

• In the United States, where inflation is projected to remain above 2% during the next two years, the Federal Reserve is now beginning to reduce net asset purchases, with a gradual tapering continuing throughout 2022. Gradual successive increases totalling 150 basis points in the Federal Funds rate are projected to follow by the end of 2023, beginning in the latter half of 2022.

• In the euro area, where underlying inflation pressures are expected to rise more slowly, and to remain close to but under 2%, the ECB is projected to continue reducing net asset purchases gradually, with policy rates unchanged in the next two years.

• In Japan, the Bank of Japan is expected to keep policy rates unchanged throughout the projection horizon. Inflation is projected to remain at or below 1% through 2022-23, in line with the recent assessments of the Policy Board members of the Bank of Japan.20

• Policy rate increases are also anticipated in the United Kingdom and Canada. In the United Kingdom, where global cost pressures have continued to affect consumer goods prices, the Bank of England is expected to raise its Bank Rate to 0.5% in the projection period. In Canada, the first increase in policy interest rates is assumed to take place in the second half of 2022, with three further rate rises, totalling 100 basis points, to follow in 2023.

These policy steps should remain state-dependent, so that central banks can respond to unexpected developments in activity and labour markets, financial conditions and the broader inflation outlook. The projected pace of mild policy normalisation over 2022-23 implies a continued accommodative stance to support the recovery. Central banks should be ready to tighten policy more rapidly if the recovery is faster than expected or if signs of more broad-based or durable inflation pressures emerge, as has already been the case in a number of smaller, open advanced economies.

Macroprudential policy should be deployed where necessary to mitigate financial market risks. In many countries, macroprudential tools such as the countercyclical capital buffer were eased aggressively in the early stages of the pandemic, as part of broader steps to provide capital relief to banks and ensure stable financial conditions. A gradual tightening should occur as the economy recovers. Addressing pockets of risks in housing markets will be particularly critical. House prices have continued to increase rapidly, partly due to a prolonged period of very low interest rates, and valuations are at record highs in many countries (see above). The ensuing reduced affordability for many households could also add to wage pressures in some countries. In addition, soaring house prices could lead to a build-up of vulnerabilities, both in the corporate (real estate) and household sectors. In the latter, measures such as loan-to-value or debt-service coverage ratios are essential to prevent excessive risk-taking and should be tightened if necessary. These efforts should be complemented by a thorough stress testing of banking sector resilience to potential fluctuations in property markets. In parallel, steps are needed to address vulnerabilities in non-bank financial intermediaries, account for a rising share of real estate financing over the last decade.

The monetary policy stance has already been tightened substantially in some major emerging-market economies like Brazil and Mexico, amidst rising inflationary pressures from higher food and energy prices, past currency depreciations and supply-chain disruptions (Figure 1.35). Pent-up demand owing to government transfers has increased services inflation in Brazil, and a high integration in global value chains has passed the effect of supply-chain bottlenecks on to domestic prices in Mexico. In contrast, monetary policy is projected to remain broadly accommodative in South Africa, with only a modest increase in the key policy rate, enabled by a strong currency and well-anchored inflation expectations. After remaining unchanged this year, key policy rates are also projected to rise only slowly in 2022 in India and Indonesia, where spare capacity remains sizeable. Lower food price inflation in India, due to the normalisation of supply chains in agricultural production, and both well-anchored inflation expectations and a limited pass-through of global prices into administered prices in Indonesia have held down domestic price pressures in these countries. In China, financing costs have been eased by a reduction in reserve requirements on banks and no monetary policy rate increase is projected, with the pass-through to headline inflation from currently rising producer prices remaining very limited. Following recent policy interest rate decreases in Turkey, further currency depreciation is adding to monetary policy challenges, and inflation is projected to be well above the target range of the central bank for some time in the absence of moves to tighten monetary policy. Asset purchase programmes introduced at the onset of the pandemic have been discontinued in almost all major emerging-market economies, with the exception of Indonesia. Clearly communicating the modalities and exit strategies of those interventions would ensure that they ease financial conditions effectively without de-anchoring inflation expectations (Mimir and Sunel, 2021).

Figure 1.35. Many emerging-market economy central banks have tightened monetary policy

Change in the main policy interest rate

Note: Based on information as of 24 November 2021.

Source: Bank for International Settlements; Central banks; and OECD calculations.

 StatLink https://stat.link/3eczjp

Fiscal policy support should remain flexible and contingent on the state of the recovery

After a strongly expansionary fiscal stance in 2020-21, governments now face complex policy challenges. The pace of withdrawal of pandemic-related measures should balance the preservation of necessary support to the recovery with the need to avoid hampering required resource reallocation. Over a longer horizon, steps will be needed to ensure the sustainability of the public finances and support the transition to carbon neutrality. These challenges have important implications for the composition of public expenditure and revenue, and call for strengthened fiscal frameworks to convey clear policy guidance to markets and the public opinion.

Fiscal policy has remained supportive this year, largely due to the continued implementation of measures announced in 2020 and early 2021. Discretionary fiscal easing, as approximated by the change in the underlying primary balance, a conventional though uncertain measure of the fiscal stance, is estimated to be 0.8% of potential GDP in the median OECD economy in 2021 (Figure 1.36), and is often larger across Europe. This partly reflects the beginning of the implementation of the Next Generation EU (NGEU) recovery plans.

Fiscal projections for 2022-23 are conditional on announced government measures and OECD assessments of current plans (Annex 1.A). In the median OECD economy, the underlying primary balance is estimated to increase by 0.8% of potential GDP in 2022, and by 0.6% in 2023:

• In the United States, most crisis-related fiscal support has already been withdrawn. The projections take account of the recently legislated infrastructure bill and assume that an additional budget reconciliation bill currently being negotiated comes into force, the two bills jointly adding around 0.3% of GDP of spending net of tax increases in both 2022 and 2023. Measures in these bills mainly come from the proposals under the American Jobs Plan (decarbonisation, infrastructure and research and development) and American Families Plan (support for low-income families), and the ensuing additional spending will be mostly tax-financed. Even with this additional spending, there may be a substantial tightening of the fiscal stance in the next two years, with the underlying primary balance projected to improve by more than 5% of potential GDP over 2022-23.

• In the EU, the assessment and approval of national recovery and resilience plans is proceeding swiftly, and initial disbursements of NGEU grants to recipient countries have begun. The implementation of plans, which is to take place until 2026, is projected to be significantly frontloaded in the two largest recipients of support, Italy and Spain, who are both expected to absorb 60% or more of the respective national grant envelopes over 2021-23 (the median absorption rate is estimated at 54%). In the euro area, stimulus from NGEU implementation, with absorbed grants projected to exceed 0.5% of euro area GDP in both 2022 and 2023, is not projected to prevent a significant increase in the underlying primary balance (1% of potential GDP over those two years combined).

Figure 1.36. The fiscal stance is normalising as crisis-related support is withdrawn

Change in the underlying primary balance, in per cent of potential GDP

Note: Vertical lines indicate the medians.

Source: OECD Economic Outlook 110 database; and OECD calculations.

 StatLink https://stat.link/xloye3

• In Japan, the government which took office in October announced a new economic package, which will be approved in the FY 2021 first supplementary budget and the FY 2022 initial budget. The package is likely to include cash benefits for vulnerable households and measures to support businesses hardest hit by COVID-19 and sectors that can help strengthen supply chain resilience, as well as policies aimed at longer-term growth and redistribution. As a consequence, a substantial loosening of the fiscal stance is projected in 2022 (2 per cent of potential GDP), boosting economic activity. With the recovery projected to be well on track, considerable consolidation efforts are expected in 2023.

Underlying primary budget balances are projected to increase in 2022 and 2023 in a large majority of economies (Figure 1.36, Panels C and D). The withdrawal of crisis-related support measures, which is now under way in many countries and assumed to proceed in 2022, is the main driver of the sizeable increase of underlying balances in 2022 and partly explains their additional improvement in 2023. Projected fiscal developments also reflect, especially in 2023, discretionary deficit-reduction measures, which are likely to weigh on activity.

With the recovery in activity and declining deficits, public debt ratios are projected to stabilise as of 2021, and even start declining in some countries (Figure 1.37). Nonetheless, debt ratios in 2023 are likely to exceed 2019 levels considerably (by 14 percentage points in the median OECD economy), and will need to be adjusted over the medium term given future demands on the public finances from long-term trends such as ageing populations (Guillemette and Turner, 2021). At the same time, debt-servicing burdens remain low due to very low interest rates. This provides room, while interest rates remain low, for additional fiscal support where needed, including measures to accelerate the transition towards climate neutrality.

The removal of pandemic-related support will need to be gradual to minimise contractionary impacts on activity and preserve space to sustain higher levels of public investment (see below). Credible fiscal frameworks that provide clear guidance about the medium-term path towards sustainability, and the likely policy changes along that path, would help to maintain market confidence and public support. Enhancing ownership and reconciling sustainability with sufficient counter-cyclicality should guide efforts to reform fiscal governance.

Reforming the composition of public finances can have important payoffs in terms of stronger medium and long-term growth, which should be the mainstay of public debt sustainability by enabling a gradual reduction of debt-to-GDP ratios. After the global financial crisis, developments in public finance composition were often adverse; for instance, many countries cut public investment. The policy response over the next few years appears likely to be more favourable. Many countries are expected to reallocate budget resources towards public investment in 2021-23, though on a relatively modest scale (Box 1.5). On the revenue side, compositional changes seem more muted, possibly suggesting an unexploited potential for tax reforms that promote equity, growth and environmental sustainability (OECD, 2021e). Priorities vary across countries, but often include reducing labour taxation on low-wage earners, increasing or broadening environmental taxes and property taxes, and expanding tax bases (OECD, 2021f).

Figure 1.37. General government gross financial liabilities

Source: OECD Economic Outlook 110 database; and OECD calculations.

 StatLink https://stat.link/pco2iz

Reforms to public spending and revenue can also foster the transition to a low-carbon economy. Higher public investment and support for innovation are essential to strengthen incentives for private low-carbon investment and to meet climate-neutrality targets, and their benefits will often be enhanced by coordination among countries — examples include cross-border grid interconnections, offshore grids and recharging infrastructure interoperability (OECD, 2021g). At the same time, investments that are inconsistent with the transition to a low-carbon economy should be phased out.21 On the revenue side, a necessary step is to increase the effective price of carbon emissions. Possible policy tools include carbon taxes, cap-and-trade schemes and the elimination of fossil fuel tax exemptions and reduced rates. It will be important to dedicate most of the ensuing additional revenues to the provision of green infrastructure and targeted support to vulnerable households and firms in the energy transition. However, relatively few countries envisage significant policy action on this front in 2022-23.22

Fiscal positions have started to improve in emerging-market economies as output recovers, but the outlook for the public finances is uneven. The recovery has been very gradual in some countries, especially lower-income countries, which reduces the feasibility, and indeed the appropriateness, of a quick reduction in deficits. In addition, rising financing costs have made it more difficult for many countries to reduce public debt-to-GDP ratios. Higher commodity prices have had heterogeneous impacts, bolstering fiscal revenues in commodity exporters like Argentina, Chile and South Africa, but hampering fiscal consolidation — including through a tightening of financing conditions — in commodity importers like India and Turkey.

Provided market access is preserved, delaying steps to ensure the long-term sustainability of the public finances may be justified until emerging-market and developing countries can reach a robust recovery path enabled by an improved health outlook. However, once that has been achieved, those steps should accelerate to preserve market confidence and create buffers. For example, in China, a number of tax exemptions are being phased out, which is projected to strengthen fiscal balances. In contrast, despite being well into the recovery phase from the pandemic, increased political uncertainty and imminent elections are delaying fiscal consolidation efforts in some countries, like Brazil and Chile. Countries with limited fiscal space face difficult choices in ensuring fiscal sustainability, and should pursue fiscal reforms that enhance medium-term growth. In Colombia, a recently legislated tax and spending reform goes some way towards halting the rise in debt, but does not go far enough in reducing tax expenditures and switching to a more growth-friendly tax mix.

Box 1.5. Recent and prospective changes in the composition of public finances across the OECD

The COVID-19 crisis has given rise to exceptional fiscal support measures, including spending on healthcare and vaccination, subsidised job retention schemes, social transfers, tax and social security contribution deferrals and moratoria on various charges and fees. As the recovery strengthens, governments face the challenge of replacing emergency support with measures that foster sustainable and equitable growth. This box tracks the recent and expected evolution of public spending and revenue composition in 33 OECD economies over 2019-23. It also compares the projected paths of public investment with model simulations based on previously estimated growth effects from fiscal expenditure reallocation.

Results highlight common patterns across countries. As a reflection of fiscal expansion, both the frequency and size of spending item increases (relative to 2019 levels) generally exceed those of revenue items (Figure 1.38). On the expenditure side, the residual category (denoted E_OTH, which is generally small) increases in most countries — by close to 2% of potential GDP in the median country — during the emergency period of 2020-21 (Figure 1.38, Panel A). Most of this spending shift is explained by the large increase in various subsidies to respond to the pandemic, which were on average 1.6% of potential GDP higher in 2020 relative to 2019.1 As the recovery progresses, this residual spending category is expected to gradually decline. At the same time, public investment (IGAA), which understandably was not the key priority when the pandemic struck, is expected to increase in 2021-23 relative to 2019 levels in a growing number of countries, albeit by a relatively modest 0.3-0.4 per cent of potential GDP. Finally, the persistence of generally low interest rates until 2023 should lead to a continued decline in implicit interest rates on public debt, and low government interest payments.

On the revenue side, as household incomes have been preserved in many countries (largely thanks to public expenditure, as discussed above), taxes on households (TYHA) rose in 2020, an increase expected to wane gradually in subsequent years (Figure 1.38, Panel B). A similar pattern, though somewhat mitigated, holds for social security contributions (SSRGA), as job retention schemes kept many workers on the payroll of employers. In a low interest rate environment, the property income received by governments (YPERG) has become less important. In contrast, the residual revenue category (R_OTH) is rising as a share of potential GDP in a majority of countries in 2022-23, mainly due to receipts of NGEU grants by EU member states.

Figure 1.38. The fiscal spending and revenue mixes are evolving

Share of countries (in per cent) with an increase relative to 2019 in the ratios of cyclically-adjusted budget items to potential GDP

Note: The spending items considered are public consumption (CGAA), social security benefits (SSPGA), public investment (IGAA), property income payments (YPEPG) and a residual category (E_OTH) which includes subsidies. The revenue items are social security contributions (SSRGA), direct taxes on households (TYHA) and on firms (TYBA), indirect taxes (TINDA), property income received (YPERG) and a residual category (R_OTH) which includes, inter alia, sales of goods and services and capital transfers received. The sample includes all 38 OECD countries except for Chile, Colombia, Costa Rica, Mexico and Turkey, due to missing data. Figures for 2021-23 reflect OECD projections. Numbers in parentheses in the vertical axes denote the median ratio of the respective budget item to potential GDP in 2019 across the 33 OECD countries considered. Numbers to the right of bars show the rounded median increase in the respective budget item in percentage points of potential GDP, computed across the subset of countries recording an increase. The OECD cyclical adjustment methodology (Price et al., 2015) is applied to those items deemed sensitive to the business cycle – i.e. all revenue items except for property income and the residual item, and, on the expenditure side, social benefits.

Source: OECD Economic Outlook 110 database; and OECD calculations.

 StatLink https://stat.link/fjw2sp

Increasing public investment would strengthen the recovery from the pandemic

Under sound governance arrangements, public investment generates positive externalities, increases the productive capital stock and mitigates underinvestment by the private sector due to market failures (Fournier, 2016; Pain et al., 2018). OECD estimates show that shifting the composition of public expenditure towards public investment while keeping total spending unchanged (henceforth referred to as “spending-neutral” shifts) can increase potential growth (Fournier and Johansson, 2016; Cournède et al., 2018). These estimates can be used to compute how much spending reallocation towards public investment would be needed to increase potential GDP per capita by a certain amount in the medium and long run – for instance, to achieve a 2% increase by 2030. This would correspond to an improvement of approximately 0.2 percentage points per annum in the annual growth of potential GDP per capita in the coming decade in the typical country (Figure 1.39).2

Figure 1.39. Projected public investment increases are welcome but often too modest

Changes in the average annual public investment-to-potential GDP ratio

Note: Green bars show the difference between the average public investment-to-potential GDP ratio in 2021-23 and that in 2019, based on the OECD projections. The red line with markers shows the spending-neutral increase in the average annual public investment-to-potential GDP ratio over 2021-30 that is required to boost potential GDP per capita by 2% in 2030. The growth effects of spending reallocation towards public investment are estimated keeping government size unchanged, and thus imply proportional decreases in spending on other budget items, whose growth effects, positive or negative, are controlled for. “OECD” shows the unweighted average of countries with available data that are included in the figure.

Source: OECD Economic Outlook 110 database; Fournier and Johansson, 2016; Cournède et al., 2018; and OECD calculations.

 StatLink https://stat.link/im1s6z

These simulation results are stylised and illustrative, inter alia because the required compositional shifts reflect average effects across countries that abstract from country-specific features, such as the existing level of the public capital stock and potentially diminishing returns to further investment. Nonetheless, the results suggest that there may be a case for larger public investment increases than those currently being envisaged until 2023. In particular, OECD countries are projected to increase their public investment-to-potential-GDP ratio by 0.4 percentage points on average in 2021-23, which is lower than the average annual amount that would be required to boost potential GDP per capita by 2% in 2030.3 The few countries with ambitious projected increases in public investment in the near future tend to be among the major recipients of NGEU grants, such as Greece, Spain and Portugal.

← 1. In contrast, the other expenditure items of the residual category have on average either remained broadly stable or, in the case of capital transfers, recorded a moderate increase (an average of 0.5 per cent of potential GDP).

← 2. Annual potential GDP per capita growth has been subdued since the global financial crisis, at about 1.1% in the median OECD country over 2010-19. This is about 0.9 percentage points per annum lower than during in the decade before the global financial crisis.

← 3. If countries further increase investment later in this decade, the gap to the required increase for a 2% gain in potential GDP per capita will be reduced or even fully eliminated, even though that gain may be reached with some delay relative to 2030.

Structural priorities to enhance resilience and restore dynamism

Economic policy efforts for most of the past two years have largely been focussed on coping with the effects of the pandemic. As those effects dissipate, it becomes increasingly possible to switch the focus away from near-term rescue efforts to longer-term priorities, with macroeconomic policy support accompanied by effective and well-targeted reforms. A key priority is the need to enhance economies’ resilience, which has been highlighted by the onset of the pandemic and the recovery from it. It is also necessary to tackle longstanding structural challenges such as digitalisation and the transition to zero net carbon emissions. Continued income support for the poorest households, enhancing activation and skill acquisition, and strengthening economic dynamism by tackling barriers to market entry and exit, will maintain demand, improve labour market opportunities and help to foster productivity-enhancing reallocation. Environmental policy challenges differ across countries, but early signals about the future trajectory of carbon prices, greater public sector support for innovation and investment, and ensuring policies are communicated clearly and accompanied by redistribution where necessary will be of critical importance everywhere.

In virtually all economies, the pandemic provoked a deep recession and major structural changes. It remains unknown how long-lasting the effects will be, but the shocks have underlined the importance of policies to facilitate the reallocation of resources between activities. Initial evidence from three OECD countries suggests that the tendency for high productivity firms to expand and low productivity firms to contract – which propels aggregate medium-term productivity growth – remained intact in the early phase of the pandemic, implying that job retention schemes did not necessarily distort firm dynamics (Andrews et al., 2021). Some evidence from Australia suggests that distortive effects may build over time, however, pointing to a danger of retaining job retention schemes longer than is needed.

The need to accommodate the shifts in activity arising from the pandemic at minimum cost, and address the potential long-term costs from the disruption to schooling during the pandemic, adds to the pre-COVID challenges requiring structural policy action. Many OECD economies were characterised by high and often growing inequalities of income and/or wealth, and all were faced with a host of underlying challenges. Governments need to seize the opportunity at a time when macroeconomic policies are supportive and demand is rising quickly to accelerate reforms. This will ensure that the extraordinary support mobilised to combat the current crisis, including plans to boost public investment, also advances longer-term objectives.

A key priority is to support people by maintaining adequate income support for those most in need whilst enhancing reforms that improve labour market opportunities, alleviate skill shortages and foster reallocation. Many countries face particularly important challenges in reforming their labour markets. Nearly two-fifths of the recommendations made by the OECD in its 2021 Going for Growth report (OECD, 2021f) are for reforms to improve the functioning of labour markets and enrich job prospects. Substantial additional investments are needed in activation and enhanced vocational education and training to boost skills and create new opportunities for displaced workers, lower-skilled workers, youth and those still on reduced working hours. Targeting training courses to individuals and delivering them in a flexible manner, including through online teaching tools, can help to allow training to be combined with part-time work and irregular work schedules.

There is also a need, highlighted by the pandemic, for governments to improve the codification and enforcement of labour standards for migrant workers. The vulnerability associated with an inability to change employers will remain a salient concern, and countries should consider how to make worker protections a part of their long-term labour migration structure. In addition, countries would generally benefit from raising immigration limits to make up for at least some of the foregone immigration during the pandemic. Canada is one country that has done this, raising its 2021 target by 18%.

As the recovery progresses, the focus for policy should increasingly move towards improving the prospects for sustainable and equitable growth. This includes additional public investment in health, digital and low-carbon infrastructure and changes in the composition of taxation. Improving broadband connectivity, helping firms develop online business models and enhancing digital skills are all areas in which further reforms would accelerate the adoption of digital technologies. Well-designed infrastructure investment projects, including expanded and modernised electricity grids and spending on renewables (coordinated across countries where relevant), and projects with shorter payback periods, such as more energy-efficient buildings and appliances, can also serve the twin objectives of closing employment gaps and achieving climate-related goals.

Climate change and the degradation of the environment are a key challenge that requires actions across a wide range of policy areas. Global carbon dioxide (CO₂) emissions must decline by 45% (compared to 2010 levels) by 2030 and reach at least net-zero globally by 2050 to hold the increase in global average temperature to 1.5°C above pre-industrial levels. To achieve this goal, the scale of annual CO₂ emissions reductions through 2030 would need to be of a similar order of magnitude to that seen in 2020, when emissions declined largely due to drastic restrictions on mobility. The latest United Nations Production Gap report (United Nations, 2021) finds that governments plan to produce more than twice the amount of fossil fuels in 2030 than would be consistent with limiting long-term warming to 1.5°C above pre-industrial levels. Analysis by the International Energy Agency (IEA) indicates the pledges made for the COP26 meeting in Glasgow, while a positive step, secure less than half the reduction in emissions needed to be consistent with the Net Zero Scenario (Figure 1.40). In particular, concerning the crucial decade ahead, the climate pledges made globally leave a 70% gap in the amount of emissions reductions needed by 2030 to keep 1.5°C within reach (Birol, 2021). 

Figure 1.40. CO2 emissions in World Energy Outlook scenarios over time, 2000-2050

Source: International Energy Agency (IEA), CO2 emissions in World Energy Outlook scenarios over time, 2000-2050, IEA, Paris https://www.iea.org/data-and-statistics/charts/co2-emissions-in-world-energy-outlook-scenarios-over-time-2000-2050-and-corresponding-global-temperature-rise-in-2100

 StatLink https://stat.link/1vlp8d

Establishing transparent and participative institutions and governance frameworks would facilitate the design and implementation of decarbonisation strategies and bolster public acceptability. Developing data and indicators that are comparable across countries and take into account current and future trends in emissions would allow for a more systematic evaluation of countries’ performance and enable progress to be monitored. International benchmarking efforts, such as those proposed by the OECD International Programme for Action on Climate, provide a useful platform in this direction.

Emission pricing and standards and regulations can complement each other, with emission pricing helping to speed up the deployment of low-carbon technologies and products as encouraged or mandated by standards and regulations. In addition, complementary policies – for example, to offset any regressive distributional implications of mitigation policies, to reskill workers in transition, or to facilitate green infrastructure investment – are key, as is engagement with stakeholders on the design of climate policy packages.