copy the linklink copied!5. Health

Health is about being and feeling well: a long life unencumbered by physical or mental illness, and the ability to participate in activities that people value. Average life expectancy at birth in OECD countries is 80.5 years, and two-thirds of adults report good health. Suicide, acute alcohol abuse and drug overdose cause 2% of all deaths. In European OECD countries, 6% of adults recently experienced depressive symptoms. Since 2010, life expectancy has increased almost everywhere, but is showing signs of plateauing in some countries. Trends in perceived health, suicide and substance abuse deaths diverged between countries. Women live longer than men, but report worse health and higher rates of depressive symptoms. Four times more men than women die from suicide and substance abuse, although female deaths from these causes have risen in more than one-third of OECD countries since 2010. There are large education- and income-related inequalities in health.

    
copy the linklink copied!
Figure 5.1. Health snapshot: current levels, and direction of change since 2010
Figure 5.1. Health snapshot: current levels, and direction of change since 2010

Note: The snapshot depicts data for 2017, or the latest available year, for each indicator. The colour of the circle indicates the direction of change, relative to 2010, or the closest available year: consistent improvement is shown in blue, consistent deterioration in orange, no clear trend in grey, and insufficient time series to determine trends in white. For each indicator, the OECD country with the lowest (on the left) and highest (on the right) well-being level are labelled, along with the OECD average in black. For full details of the methodology, see the Reader’s Guide. * for perceived health signifies a different reporting scale, which may lead to an upward bias in their reported estimates.

Source: OECD Health Status (database), http://stats.oecd.org/Index.aspx?DataSetCode=HEALTH_STAT; Eurostat’s European Health Interview Survey (database), https://ec.europa.eu/eurostat/web/microdata/european-health-interview-survey.

copy the linklink copied!Life expectancy at birth

Newborns in more than two-thirds of OECD countries can expect to live beyond 80 years (80.5 years on average for the OECD as a whole), and up to 84.2 in Japan (Figure 5.2). Life expectancy has increased in all OECD countries over the last few decades and was over ten years higher in 2017 than it was in 1970 (OECD, 2019[1]). Compared to 2010, average life expectancy has increased by about 1 year and 2 months (1.5%). Yet growth has slowed in some countries: for Iceland, Germany, Greece and the United Kingdom, life expectancy is plateauing, with gains of less than 9 months between 2010 and 2017. In the United States, already below the OECD average at 78.6 years, net gains in life expectancy over this time have been nil, after a temporary decrease over 2014-17. The causes of the slowdown in life expectancy gains are multifaceted: Improvements in heart disease and stroke have slowed as populations age and levels of obesity and diabetes rise, a comparatively large number of people died from influenza and pneumonia in the recent decade, and drug-related accidental poisoning rose in some countries in the context of the opioid crisis (OECD, 2019[1]; Raleigh, 2019[2]). But there is also good news: many countries with comparatively lower levels of life expectancy are converging towards the OECD average. For example, life expectancy has risen by more than 2 years since 2010 in Chile, Estonia, Korea and Lithuania.

copy the linklink copied!
Figure 5.2. Longevity gains since 2010 have slowed in some countries, and are often larger in countries below the OECD average level
Life expectancy at birth, years
Figure 5.2. Longevity gains since 2010 have slowed in some countries, and are often larger in countries below the OECD average level

Note: The latest available year is 2016 for Chile. The earliest available year is 2011 for Belgium and Switzerland, 2012 for Hungary and Luxembourg, 2013 for Turkey and 2014 for the Russian Federation.

Source: OECD Health Status (database), http://stats.oecd.org/Index.aspx?DataSetCode=HEALTH_STAT.

 StatLink https://doi.org/10.1787/888934081283

copy the linklink copied!Perceived health

On average, between 6 and 7 out of 10 people in OECD countries say their health is in good shape (Figure 5.3). However, there are notable country differences: in Asian and eastern European OECD members, as well as in Portugal, fewer than 60% of adults view their health as good. By contrast, more than 80% in Australia, Canada, Ireland, New Zealand, Switzerland and the United States do so (though differences in the way survey questions are phrased in some of these countries might bias results upwards). While the OECD average has remained relatively stable, trends since 2010 have diverged between countries. Perceived health has improved most in Slovenia (5.7 percentage points), and declined most in Lithuania (-6.5) and Korea (-8.1).

copy the linklink copied!
Figure 5.3. Around two-thirds of people in OECD countries say their health is good
Share of the population aged 15 and over reporting “good” or “very good” health, percentage
Figure 5.3. Around two-thirds of people in OECD countries say their health is good

Note: The latest available year is 2016 for Iceland and Japan. The earliest available year is 2011 for Australia and 2012 for New Zealand. Respondents in European OECD countries are generally aged 16 years+, those in Australia, Canada, Costa Rica, Chile, Japan, Korea, Lithuania, New Zealand, Turkey and the United States 15 years+, and those in Israel 20 years+. Australia, Canada, Chile, Israel, New Zealand and the United States (shown in grey) use a different reporting scale, which leads to an upward bias in the results. The OECD average excludes Mexico, due to a lack of available data, Chile, due to a break in the series, and Australia, Canada, Israel, New Zealand and the United States, due to differences in methodology.

Source: OECD Health Status (database), http://stats.oecd.org/Index.aspx?DataSetCode=HEALTH_STAT.

 StatLink https://doi.org/10.1787/888934081302

copy the linklink copied!Deaths from suicide, acute alcohol abuse and drug overdose

Fatalities from suicide, acute alcohol abuse and drug overdose have recently been coined as “deaths of despair” (Case and Deaton, 2017[3]). On average, 14.8 people per 100 000 in OECD countries die from such causes, which is only a small share (1.8%) of overall deaths (Figure 5.4). Nevertheless, these deaths represent an important measure of severe mental illness and addiction among the population (OECD, 2019[4]). Slovenia, Lithuania and Latvia, as well as Korea and Denmark, record the highest death rates from suicide and substance abuse in the OECD, above 20 per 100 000 population. Among these, deaths of despair are mainly from suicide in Korea and Lithuania, whereas fatalities from acute alcohol abuse represent at least a third of overall deaths of despair in Latvia, Denmark and Slovenia (Figure 5.4). By contrast, overall rates are very low in Turkey (2.6), Greece (4.2) and Colombia (5.2). Yet these estimates should be interpreted with some caution, since death registries are likely to underrepresent the phenomena due to different reporting practices and stigma (Box 5.1).

Since 2010, deaths from suicide (the most common form of deaths of despair, Figure 5.5) and substance abuse have fallen in a third of OECD countries, driven mainly by reductions in suicides. Some of the countries with the greatest challenges have made the most progress: Hungary, Japan and Korea reduced these fatalities by over 25%, Estonia by 23% and Lithuania by 15%. The situation worsened elsewhere: since 2010, deaths of despair increased by 16% in the United States, 18% in Slovenia (with the highest level in the OECD) and 30% in the Netherlands. In these three countries, deaths from both acute alcohol abuse and drug overdose rose substantially.

copy the linklink copied!
Figure 5.4. Deaths of despair have fallen in some of the OECD countries where rates are highest, but increased elsewhere
Combined deaths from suicide, acute alcohol and drug use abuse, per 100 000 population (age-standardized)
Figure 5.4. Deaths of despair have fallen in some of the OECD countries where rates are highest, but increased elsewhere

Note: For each of the causes of death, the closest available datapoint to 2016 and 2010 is considered separately: The latest available year is 2015 for Canada, Denmark, France, Ireland, Italy, Latvia and South Africa (all types of deaths), as well as Slovenia (suicide and acute alcohol abuse), Colombia (acute alcohol and drug abuse), Brazil, Estonia, Greece, Iceland and Japan (drug abuse); 2014 for Costa Rica, the Slovak Republic, New Zealand (all types of deaths), Brazil (acute alcohol and drug abuse) and the Russian Federation (suicides); and 2013 for Korea and Slovenia (drug abuse). The earliest available year is 2011 for Ireland (acute alcohol abuse) and Estonia (drug abuse), 2009 for Iceland (drug abuse), 2008 for Slovenia (drug abuse) and 2006 for Luxembourg (drug abuse). The OECD average excludes Greece, Ireland and the Slovak Republic, for which data on acute alcohol and drug abuse is missing for the earliest available year. Data for the Russian Federation refer to suicides only.

Source: OECD Health Status (database), http://stats.oecd.org/Index.aspx?DataSetCode=HEALTH_STAT.

 StatLink https://doi.org/10.1787/888934081321

copy the linklink copied!Depressive symptoms

Data on self-reported depressive symptoms are available only for European OECD countries, where, on average, 6% of adults experienced a range of depressive symptoms in the past two weeks (e.g. having little interest in doing things, feeling tired, overeating or having no appetite) (Figure 5.6). Slightly more people, 8% on average, self-report having suffered from chronic depression (the most common mental disorder after anxiety disorder in the EU) in the past year (OECD/EU, 2018[5]).

copy the linklink copied!
Figure 5.5. Suicide is the most common death of despair, followed by alcohol-related fatalities
Deaths from suicide, acute alcohol abuse and drug overdose, per 100 000 population, 2016
Figure 5.5. Suicide is the most common death of despair, followed by alcohol-related fatalities

Note: See the note of Figure 5.4 for reference years and further details. Data for the Russian Federation refer to suicides only.

Source: OECD Health Status (database), http://stats.oecd.org/Index.aspx?DataSetCode=HEALTH_STAT.

 StatLink https://doi.org/10.1787/888934081340

copy the linklink copied!
Figure 5.6. 6% of adults in European OECD countries recently experienced depressive symptoms
Share of respondents reporting depressive symptoms in the past two weeks, percentage, 2014
Figure 5.6. 6% of adults in European OECD countries recently experienced depressive symptoms

Source: Eurostat’s European Health Interview Survey (database), https://ec.europa.eu/eurostat/web/microdata/european-health-interview-survey.

 StatLink https://doi.org/10.1787/888934081359

copy the linklink copied!Health inequalities: gaps between population groups

Men live shorter lives and suffer more deaths of despair, but report better health and fewer depressive symptoms than women

Life expectancy at birth is higher for women (83.2 years, on average) than for men (77.9 years) in all OECD countries. Conversely, 70% of men report their health to be good, but only 66% of women do, on average. These gender gaps vary in size across countries, but the direction remains consistent in almost all cases (Figure 5.7). Eastern European countries are furthest from gender parity on both measures.

copy the linklink copied!
Figure 5.7. Women live longer than men, but perceive their overall health to be worse
Gender ratio for life expectancy at birth and percentage of adults reporting “good” or “very good” health, 2017
Figure 5.7. Women live longer than men, but perceive their overall health to be worse

Note: The gender ratio is calculated by dividing average values for women by average values for men. Thus, values above 1 always indicate better outcomes for women, and values below 1 always indicate better outcomes for men. See the notes of Figure 5.2 and Figure 5.3 for reference years and further details. The OECD average for perceived health excludes Mexico, due to a lack of available data.

Source: OECD Health Status (database), http://stats.oecd.org/Index.aspx?DataSetCode=HEALTH_STAT.

 StatLink https://doi.org/10.1787/888934081378

Throughout OECD countries, men are much more likely to die from suicide, acute alcohol abuse or a drug overdose – on average, almost 4 men for every woman (Figure 5.8). This gender gap is largest in Poland, at 8.2. Even in the country with the smallest gender gap (Luxembourg), the rate of deaths among men is double the rate for women.

The size of the gender gap in deaths of despair has narrowed in 20 OECD countries since 2010. In more than half of these, this has been driven by a higher or stagnant female death rate alongside fewer male deaths. Overall, female deaths from suicide or substance abuse increased in more than one-third of OECD countries (14) since 2010. Nevertheless, in two of the most unequal countries (Iceland, Lithuania), the gap between the sexes widened further, as deaths among women decreased at a faster pace than those for men.

In the European OECD countries where data are available, more women (8%) than men (5%) have experienced recent depressive symptoms (OECD/EU, 2018[5]).

copy the linklink copied!
Figure 5.8. Gender gaps in deaths of despair have narrowed, but many more men than women continue to die from suicide and acute substance abuse
Gender ratio for combined deaths from suicide, acute alcohol abuse and drug overdose, per 100 000 population
Figure 5.8. Gender gaps in deaths of despair have narrowed, but many more men than women continue to die from suicide and acute substance abuse

Note: Gender ratios are calculated by dividing average values for men by average values for women. Thus, values above 1 indicate higher relative deaths of despair rates for men, and those below 1 indicate higher relative deaths of despair rates for women. See the note of Figure 5.4 for reference years and further details. In addition, for women, the latest available year is 2014 for Colombia and Japan and 2015 for Portugal and Turkey. The earliest available year for women is 2011 for Israel, Japan, Portugal and Turkey. The OECD average excludes Estonia, Greece, Iceland, Ireland, Korea, Latvia, Poland and the Slovak Republic due to a lack of available data by gender for at least one of the time points. Data for Costa Rica and the Russian Federation refer to suicides only. Countries where women’s deaths have increased are marked in grey.

Source: OECD Health Status (database), http://stats.oecd.org/Index.aspx?DataSetCode=HEALTH_STAT.

 StatLink https://doi.org/10.1787/888934081397

People with less education and income have worse health

There are marked inequalities in life expectancy and self-reported health that are related to differences in education and income. In the 18 OECD countries for which data are available, the average gap in life expectancy at age 25 between high- and low-educated people is 7.6 years for men and 4.8 years for women (Figure 5.9). At age 65, these gaps are 3.6 and 2.6 years, respectively (Murtin et al., 2017[6]). Similarly, across all OECD countries, better educated people experience better physical and mental health: on average, 78% of those with a tertiary education say their health is good, compared to 65% of people with a secondary degree (OECD Health Status database). In European OECD countries, 4% of people with tertiary degrees versus 6% with secondary degrees have experienced recent depressive symptoms (OECD/EU, 2018[5]).

Without exception, people with higher income also report better health. On average, 79% of those in the top income quintile in OECD countries say their health is good, compared to only 60% in the bottom quintile (Figure 5.10). Eastern European countries show the largest income-related differences, with gaps in perceived health exceeding 25 percentage points. In the Czech Republic, Latvia and Estonia, income-related differences in self-reported health also widened by more than 10 percentage points since 2010.

copy the linklink copied!
Figure 5.9. Better educated people live much longer
Gap in life expectancy between people with low and high education at age 25, in years, 2011
Figure 5.9. Better educated people live much longer

Note: Data for the United Kingdom refer to England and Wales. Low education refers to people with no schooling and those with primary and lower secondary educational attainment. High education refers to people who have completed tertiary education.

Source: (Murtin et al., 2017[6]), “Inequalities in longevity by education in OECD countries: Insights from new OECD estimates”, OECD Statistics Working Papers, No. 2017/2, OECD Publishing, Paris, https://dx.doi.org/10.1787/6b64d9cf-en.

 StatLink https://doi.org/10.1787/888934081416

copy the linklink copied!
Figure 5.10. People with higher income say their health is better
Share of adults reporting “good” or “very good” health, by income quintile, percentage, 2017
Figure 5.10. People with higher income say their health is better

Note: See the note of Figure 5.3 for reference years and further details. Data for Australia, Canada, Chile, Israel, New Zealand and the United States (shown in grey) use a different reporting scale, which may lead to an upward bias in their reported estimates. The OECD average excludes them, due to differences in methodology, and Mexico, due to a lack of available data.

Source: OECD Health Status (database), http://stats.oecd.org/Index.aspx?DataSetCode=HEALTH_STAT.

 StatLink https://doi.org/10.1787/888934081435

copy the linklink copied!
Box 5.1. Measurement and the statistical agenda ahead

Health is about being and feeling well: a long life unencumbered by physical or mental illness, and the ability to participate in activities that people value. An ideal set of outcome indicators of health would provide information about good health states (feeling well; functioning well) alongside the most important diseases and conditions causing poor health, disability or death – including their prevalence, chronicity and intensity. Capturing both physical and mental aspects of health outcomes is vital – and although the latter have proved challenging to measure (particularly in international contexts), they are gaining increased recognition from policy makers, the medical community and the business world (Patel et al., 2018[7]; OECD, 2019[8]). The present chapter considers four indicators of physical and mental health (Table 5.1), as well as their distribution across the population in OECD countries.

copy the linklink copied!
Table 5.1. Health indicators considered in this chapter

Average

Vertical inequality (gap between top and bottom of the distribution)

Horizontal inequality (difference between groups, by gender, age, education)

Deprivation

Life expectancy

Number of years that a newborn can expect to live

Standard deviation of age at death

Gaps in life expectancy

n/a

Perceived health

Share of the population 16 years or over reporting “good” or “very good” health

n/a

Gaps in perceived health

Share of adults reporting “bad” or “very bad” health

Deaths from suicide, alcohol, drugs

Combined deaths from suicide, acute alcohol abuse and drug overdose, per 100 000 population (age-standardised based on the 2010 OECD population structure)

n/a

Gaps in death rates due to suicide, acute alcohol abuse and drug overdose

n/a

Depressive symptoms

Share of the population 15 years and over reporting having experienced a range of depressive symptoms in the past two weeks

n/a

Gaps in depressive symptoms

n/a

Life expectancy at birth is a summary measure of mortality rates, and refers to the number of years a child born today could expect to live based on the age-specific death rates currently prevailing. It is only an estimate of the expected life span of a given cohort, as the age-specific death rates of a particular birth cohort cannot be known in advance. The OECD computes the unweighted average of life expectancy for men and women. Education-related inequalities in longevity exist for a sub-set of countries, produced by matching census and death registry data (Murtin et al., 2017[6]).

Perceived health refers to people’s overall self-reported health status. Data are based on general household surveys or on more detailed health interviews. The indicator is based on questions such as: “How is your health in general?”, with answers usually classified as “very good”, “good”, “not very good” and “poor” – although in some non-European countries (Australia, Canada, Chile, Israel, New Zealand, the United States) different response scales are used, which may lead to an upward bias in the estimates. In the OECD Health Status database, the response categories from different surveys are rescored to fit into three broad categories of “good/very good” (all positive response categories), “fair” (neither good nor bad), “bad/very bad” (all negative response categories). Respondents are generally 16 years or over, though the specific age range varies across countries.

Deaths from suicide, acute alcohol abuse and drug overdose is an objective measure of severe mental illness and addiction. The indicator reported here is drawn from official death registries and refers to combined deaths from suicides, acute alcohol abuse and drug overdose (ICD-10 codes X60-X84,Y87.0, F10, F11-16, F18-19) per 100 000 population (standardised to 2010).

Self-reported depressive symptoms is a measure of mental (ill)health. It refers to the share of people 15 years or over who report experiencing a range of depressive symptoms in the past two weeks: little interest or pleasure in doing things; feeling down, depressed or hopeless; trouble falling or staying asleep, or sleeping too much; feeling tired or having little energy; poor appetite or overeating; feeling bad about yourself or that you are a failure or have let yourself or your family down; trouble concentrating on things, such as reading the newspaper or watching television; moving or speaking so slowly that other people could have noticed, or being so fidgety or restless that you have been moving around a lot more than usual. In line with the criteria of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), a respondent is characterised as having depressive symptoms if one of the first two items (little interest or pleasure in doing things, feeling down, depressed or hopeless) and five or more of the total list (major depression) or one of the first two items and two to four of the total list (other depressive symptoms) are reported for at least half of the reference period. The measure is limited to European OECD countries and sourced from the European Health Interview Survey.

Correlations among Health indicators

Several objective and subjective aspects of health are significantly correlated (Table 5.2). Countries where people perceive their health to be good tend to have somewhat higher levels of life expectancy (0.35) and death rates from suicide and substance abuse tend to be lower (-0.46). Depressive symptoms are not significantly correlated with the other health outcomes addressed here, suggesting this indicator provides information about mental states that is not captured through the other indicators.

copy the linklink copied!
Table 5.2. Objective and subjective measures of Health are related at the country level
Bivariate correlation coefficients among the Health indicators

Life expectancy

Perceived health

Deaths from suicide, alcohol, drugs

Depressive symptoms

Life expectancy

Perceived health

0.35**

(35)

Deaths from suicide, alcohol, drugs

0.09

(41)

-0.46***

(35)

Depressive symptoms

0.07

(24)

0.26

(24)

-0.15

(24)

Note: Values in parenthesis refer to the number of observations. * Indicates that correlations are significant at the p<0.10 level, ** that they are significant at the p<0.05 level, and *** at the p<0.01 level.

Statistical agenda ahead

While administrative data on specific disease conditions (e.g. cancer, diabetes, circulatory diseases) are available, they do not address issues of co-morbidity (i.e. the presence of different conditions affecting the same individual), which is also important for understanding people’s health-related quality of life, and the prevailing rates of disease incidence across the population (e.g. the share of people living with a serious health condition).

Life expectancy refers only to length of life, not to whether those years are spent in good health. Alternative measures of “healthy” life expectancy (based on disability weights associated with different health states, used to compute the number of years of good health that a newborn can expect to live) are not internationally comparable (except for Europe), and methods for computing disability weights remain contested. Measures of perceived health exist for the majority of the OECD, but with considerable scope to harmonise question wording and response scales.

Comparable measures of mental health outcomes are available only for European OECD countries through the European Health Interview Survey, run every 5 years. It remains challenging to identify internationally comparable mental health outcome measures at the population level (versus people diagnosed or treated by medical professionals). Measures focusing on the latter can penalise countries with good medical systems and awareness programmes, where people are more likely to seek treatment. The stigma attached to mental health may lead to underreporting, affecting cross-country comparability and the interpretation of changes in prevalence rates. Data on suicides are also likely to underrepresent the scale of the phenomenon due to stigma, and do not account for the (much higher) rate of suicide attempts.

Measures of people’s functioning (i.e. whether they can perform daily activities, including self-care) have long been recommended, e.g. by the Washington Group (Washington Group on Disability Statistics, 2016[9]). Despite international guidance (e.g. the Budapest Initiative survey module for measuring health state, prepared by the Joint UNECE/ WHO/ Eurostat Task Force on Measuring Health Status), harmonised measures are not yet available (United Nations, 2005[10]).

To make health systems more people-centred, in 2017 the OECD started an ambitious programme of work to benchmark the experiences and outcomes reported by patients themselves in the context of the PaRIS program (Patient-reported Indicator Surveys) (OECD, 2019[1]). In the future, such exercises could be extended beyond the subset of people in contact with health care to the population as a whole.

References

[3] Case, A. and A. Deaton (2017), “Mortality and Morbidity in the 21st Century”, Brookings Papers on Economic Activity, http://brookings.edu/wp-content/uploads/2017/08/casetextsp17bpea.pdf (accessed on 26 September 2019).

[6] Murtin, F. et al. (2017), “Inequalities in longevity by education in OECD countries: Insights from new OECD estimates”, OECD Statistics Working Papers, No. 2017/2, OECD Publishing, Paris, https://dx.doi.org/10.1787/6b64d9cf-en.

[4] OECD (2019), Addressing Problematic Opioid Use in OECD Countries, OECD Health Policy Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/a18286f0-en.

[1] OECD (2019), Health at a Glance 2019: OECD Indicators, OECD Publishing, Paris, https://dx.doi.org/10.1787/4dd50c09-en.

[8] OECD (2019), OECD Mental Health Performance Framework, http://oecd.org/health/health-systems/OECD-Mental-Health-Performance-Framework-2019.pdf (accessed on 14 January 2020).

[5] OECD/EU (2018), Health at a Glance: Europe 2018: State of Health in the EU Cycle, OECD Publishing, Paris/European Union, Brussels, https://dx.doi.org/10.1787/health_glance_eur-2018-en.

[7] Patel, V. et al. (2018), “The Lancet Commission on global mental health and sustainable development””, The Lancet Commissions, Vol. 392/10157, pp. 1553-1598, https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(18)31612-X/fulltext.

[2] Raleigh, V. (2019), “Trends in life expectancy in EU and other OECD countries : Why are improvements slowing?”, OECD Health Working Papers, No. 108, OECD Publishing, Paris, https://dx.doi.org/10.1787/223159ab-en.

[10] United Nations (2005), Budapest Initiative Task Force on Measurement of Health Status: Survey module for measuring health state, https://digitallibrary.un.org/record/747214 (accessed on 16 September 2019).

[9] Washington Group on Disability Statistics (2016), The Washington Group Short Set on Functioning (WG-SS), http://washingtongroup-disability.com (accessed on 16 September 2019).

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

https://doi.org/10.1787/9870c393-en

© OECD 2020

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at http://www.oecd.org/termsandconditions.