BUU, Alm.del - 2020-21 - Endeligt svar på spørgsmål 117: Spm. om de vurderinger, som OECD har lavet af hvad skolenedlukningerne under covid-19 betyder af fagligt tab og økonomisk samfundstab, til børne- og undervisningsministeren

Børne- og Undervisningsudvalget 2020-21
BUU Alm.del
Offentligt

The Economic Impacts of

Learning Losses

Eric A. Hanushek

Ludger Woessmann

September 2020

BUU, Alm.del - 2020-21 - Endeligt svar på spørgsmål 117: Spm. om de vurderinger, som OECD har lavet af hvad skolenedlukningerne under covid-19 betyder af fagligt tab og økonomisk samfundstab, til børne- og undervisningsministeren

Acknowledgements

This paper beneﬁtted from the encouragement and

support of Andreas Schleicher and from his comments

along with those of Macke Raymond, Michael Ward, and

Francesco Avvisati.

Foreword

The worldwide school closures in early 2020 led to

losses in learning that will not easily be made up for even

if schools quickly return to their prior performance levels.

These losses will have lasting economic impacts both on

the affected students and on each nation unless they are

effectively remediated.

While the precise learning losses are not yet known,

existing research suggests that the students in grades 1-12

affected by the closures might expect some 3 percent

lower income over their entire lifetimes. For nations, the

lower long-term growth related to such losses might yield

an average of 1.5 percent lower annual GDP for the

remainder of the century. These economic losses would

grow if schools are unable to re-start quickly.

The economic losses will be more deeply felt by

disadvantaged students. All indications are that students

whose families are less able to support out-of-school

learning will face larger learning losses than their more

advantaged peers, which in turn will translate into deeper

losses of lifetime earnings.

The present value of the economic losses to nations reach

huge proportions. Just returning schools to where they

were in 2019 will not avoid such losses. Only making

them better can. While a variety of approaches might be

attempted, existing research indicates that close attention

to the modiﬁed re-opening of schools offers strategies

that could ameliorate the losses. Speciﬁcally, with the

expected increase in video-based instruction, matching

the skills of the teaching force to the new range of tasks

and activities could quickly move schools to heightened

performance. Additionally, because the prior disruptions

are likely to increase the variations in learning levels within

individual classrooms, pivoting to more individualised

instruction could leave all students better off as schools

resume.

As schools move to re-establish their programmes even as

the pandemic continues, it is natural to focus considerable

attention on the mechanics and logistics of safe re-opening.

But the long-term economic impacts also require serious

attention, because the losses already suffered demand

more than the best of currently considered re-opening

approaches.

Table of contents

Acknowledgements

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2

Foreword

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3

Introduction

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5

Lost Learning during times of closed schools

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6

Economic effects of lost learning

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7

Distributional issues

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Aggregate losses in GDP across G20 nations

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Making up for learning losses

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11

Conclusions

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

Annex A. Direct evidence on the effects of closed schools

Annex B. Projection of costs from lower economic growth

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16

Notes

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17

Bibliography

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19

Figure and Tables

Figure 1•

Days of schooling lost by mid-May 2020

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 1•

Lost individual income due to Corona-induced learning loss

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 2•

Long-run loss in GDP due to Corona-induced learning loss

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 3•

Present value of lost GDP due to Corona-induced learning loss for G20 nations

. . . . . . . . . . . . . . . . . .

| The Economic Impacts of Learning Losses

Introduction

A central component of the economic development

policies of most countries has been investment in

the human capital of society. Individuals with more

skills are more productive and more adaptable to

technological changes in their economies. Nations

with more skilled populations grow faster. In many

countries, the reactions to the pandemic have,

however, threatened the long-run future of the current

cohort of students, and the harm to them from recent

events can ripple through the world’s economies in

ways that will be felt far into the future.

As the potential health threats from the COVID-19

virus began to be understood at the beginning of

2020, schools in virtually all nations closed and sent

their students home. Since then, public attention

has rightfully focused on the immediate health and

safety concerns surrounding schools, and nations are

experimenting with alternate ways to proceed with

their re-opening. Longer-run issues, however, have not

received the same attention.

The broader policy discussion has also focused

on short-run issues. Policies introduced to ﬁght the

spread of the virus consisted of various degrees of

business shutdowns and restrictions on movement

and commerce. Economic analysis has so far focused

on the near-term impact of business closures on

unemployment and on ways to provide safety nets

for individuals directly harmed, but in doing so often

leaves out consideration of longer-run issues.

Indeed, the urgency of dealing with the immediate

and obvious issues of the pandemic has pushed aside

any serious consideration of the longer-run costs of

the virus-induced school closures. There is no doubt

that the school closures in the ﬁrst half of 2020 have

resulted in signiﬁcant learning losses to the affected

cohort of students – and some of the re-opening

strategies being implemented will only further exacerbate

these already incurred learning losses. These losses

will follow students into the labour market, and both

students and their nations are likely to feel the adverse

economic outcomes.

Nobody can predict perfectly how school closures

will affect the future development of the affected

children, but past research has investigated how

school attendance and learning outcomes affect

labour-market chances and economic development.

This paper summarises the literature on the relationship

between skills and years of schooling on the one

The Economic Impacts of Learning Losses |

hand, and individual and aggregate income on the

other. In addition, it reviews previous studies of various

examples of school closures and their long-term effects

on the affected pupils.

This analysis suggests plausible ranges for the

economic impact of existing and on-going learning

losses based on existing economic research. The

typical current student might expect something on

the order of 3% lower career earnings if schools

immediately returned to 2019 performance levels.

Disadvantaged students will almost certainly see larger

impacts. And, with the forecast of further disruptions

in normal school operations, the costs only grow. For

nations, the impact could optimistically be 1.5% lower

GDP throughout the remainder of the century and

proportionately even lower if education systems are

slow to return to prior levels of performance.

These losses will be permanent unless the schools

return to better performance levels than those in 2019.

In this unprecedented period, there is of course great

uncertainty about how to develop better schools, but

two moves are suggested by existing research. First,

large differences in effectiveness of teachers have

been ubiquitous across schools. These differences are

likely to be compounded as changes in the approach

to schooling such as split shifts in schools and more

video and dispersed instruction are introduced.

Schools could improve if attention was given to using

the teachers more effectively within and across media.

Second, because the magnitude of learning losses

will differ across students, teachers will face larger

differences in preparations as the students return to

their classes. More attention to individualising the

instruction could elevate the learning for all students

and could act to ameliorate the losses from prior

closures by offering learning opportunities matched to

each student.

Lost learning during times of closed schools

For many children and adolescents, no lessons were

held in schools for at least some portion of the ﬁrst

half of 2020. Figure 1 provides estimates of days of

schooling lost to closure from an OECD/Harvard

survey conducted in mid-May 2020, a time when the

school term was not yet scheduled to end in most of

these countries. While the details of closures require

further investigation, these early estimates show

substantial losses in most countries. Additional losses

since that survey, as well as expected losses into the

future, can be expected to be much larger in most

countries.

Little is known about the effectiveness of learning at

home for the entire student population and what this

means for the development of skills. However, there

are indications from multiple countries that many

children had little effective instruction. For a signiﬁcant

proportion of pupils, learning during school closures

was apparently almost non-existent. For example, early

tracking data from an online mathematics application

used in a number of US school districts prior to

COVID-19 suggest that the learning progress of

students has suffered a strong decline during the crisis,

especially in schools in low-income areas

(Chetty et al., 2020

[1]

For Germany, a survey of parents of school children

shows that the time that children spent on school-related

activities per day was halved during the COVID-19

school-closure period, from 7.4 to 3.6 hours

(Woessmann et al., 2020

[2]

). Indeed, 38% of students

studied for school for no more than two hours per day,

74% for no more than four hours. By contrast, the time

spent with TV, computer games, and mobile phones

(passive activities) increased to 5.2 hours per day.

For children whose parents were more educated, the

decline in school activities was similar to that of other

children, although the increase in passive activities was

slightly smaller. Low-achieving students in particular

replaced learning with passive activities. Only 6% of

students had group online lessons on a daily basis,

more than half had them less than once a week.

Students had individual contact with their teachers

even less often. The standard learning tool was task

sheets that students received for weekly processing.

In sum, learning opportunities were signiﬁcantly

reduced during the school closures, and the reductions

were greatest for disadvantaged children.

Beyond the full closures already observed, schools

in many countries are also not expected to resume

normal school operations during the following school

| The Economic Impacts of Learning Losses

year. With experimentation into partial in-class work,

partial video work, asynchronous presentations, and

the various new inventions of schools, just counting past

school days lost is almost certainly underestimating the

total learning loss.

What students learn throughout the year is likely to be

signiﬁcantly less than what was seen in 2019, although

there are no good measures of hybrid learning

currently available.

Figure 1•Days

of schooling lost by mid-May 2020

Costa Rica

Moreover, it is known from many studies (such as

the analyses of skill development during the summer

holidays described in the Annex A) that learning is a

dynamic process that builds on prior learning, so that

stagnation leads to growing deﬁcits. Closed schools

not only impart less new knowledge (Oreopoulos

and Salvanes, 2011

[3]

), but also mean loss of already

acquired skills on which further learning could build

(Kuhfeld et al., 2020

[4]

Brazil

Peru

United States of America

Republic of Korea

Portugal

Norway

United Kingdom

Czech Republic

Canada

Austria

Georgia

Belgium

Sweden

Finland

Croaria

Latvia

Italy

Lithuania

South Africa

Dominican Republic

Colombia

Netherlands

Hungary

Source:

OECD/Harvard University, (2020

[5]

) Global Education Innovation Initiative at Harvard and OECD Rapid Assessment of COVID-19 Education Response.

Economic effects of lost learning

There are two related streams of long-run economic

costs that are central to this discussion. First, affected

students whose schooling has been interrupted by the

pandemic face long-term losses in income. Second,

national economies that go forward with a less skilled

labour force face lower economic growth which

subtracts from the overall welfare of society.

Much is known about the economic value of schooling

and, speciﬁcally, of cognitive skills developed through

the educational system.

Education equips people with

the skills that make them more productive at carrying

out their work tasks, particularly in modern

knowledge-based economies. Education also

provides knowledge and skills that enable people to

generate and apply new ideas and innovations that

enable technological progress and overall economic

growth.

The existing research base provides a direct means

of estimating the economic costs of learning losses.

Even though the levels of learning losses are not known

precisely, it is possible to provide estimates of the most

likely ranges of economic impact. This analysis focuses

primarily on the effects of the lack of development of

cognitive skills.

These are not the only costs. In addition, the school

closures can be expected to have numerous

consequences for the socio-emotional and

motivational development of the affected children

and adolescents. Development in these areas is

restricted by the lack of contact with classmates and

the psychological strain on families during an extended

stay in sometimes cramped housing conditions. Even

though these are not directly addressed in this analysis,

these potential deﬁcits in the development of

The Economic Impacts of Learning Losses |

Uruguay

Slovenia

Jamaica

Iceland

Estonia

Chile

Mexico

Greece

France

Spain

socio-emotional skills are likely to also negatively

impact economic potential.

Another important area that is not explicitly covered

here is early childhood education. Recent research and

the commensurate policy development have pointed

to an important role of early childhood education.

This critical development window appears particularly

important for preparing disadvantaged students for

schooling (e.g., (Heckman, 2006

[6]

)). The disruption

of this segment of the education system will likely have

lasting long-term impacts on affected child cohorts, but

it is not currently possible to incorporate this into the

estimates.

Countries vary considerably in the economic rewards

to higher skills. While workers in Singapore are

estimated to receive 50% higher income if they have

one standard deviation higher test scores, the typical

worker in Greece gains just 14% more income with one

standard deviation higher test scores. For the United

States, the comparable return to skill is 27%, and for

the average across all sampled countries it is 23%.

Importantly, these relationships provide estimates of the

impact of skill differences across the entire work life.

Table 1 displays the estimates of the percentage loss in

lifetime income for students suffering cognitive learning

losses typical of different proportions of the school

year (relying on one year of school equals one-third

standard deviations of learning). The table provides

estimates for school closures from one-quarter of a

year to a full year, and for the returns to skills found

across the sampled countries and for the United

States and the extremes of Greece and Singapore.

Looking at the losses associated with one-third of a

year closure, the pooled estimates indicate that current

students will suffer 2.6% loss in income across their

entire career. The estimated losses for this one-third

year closure exceed 3% in the US and reach 5.6% in

Singapore.

These estimates should be thought of as the lower

bound of the impact of learning losses. In addition to

the income earned, higher skills are also signiﬁcantly

linked to the likelihood of employment in the labour

market. For example, in the United States one

standard deviation of PIAAC skills is associated with a

probability of employment that is about 10 percentage

points higher (Hampf, Wiederhold and Woessmann,

2017

[8]

: Table 2). Furthermore, as discussed above, no

consideration is given to potential impacts on

socio-emotional skills.

Skills and earned income

Numerous studies show a strong association between

learned skills and the income earned in the labour

market. Consistent with the attention on learning loss,

the analysis here focuses on the impact of greater

cognitive skills as measured by standard tests on a

student’s future labour-market opportunities (Hanushek

and Woessmann, 2008

[7]

As noted, however, there is no direct evidence of the

typical loss of achievement. Various researchers have

taken different approaches to this estimation.

In order

to pin down the range of economic losses, it is useful

to begin with the simple relationship between school

years and normal learning progress. When comparing

learning gains on different tests and examinations,

these can be expressed in units of standard deviations

of the scores in the respective test populations.

rough rule of thumb, found from comparisons of

learning on tests designed to track performance over

time, is that students on average learn about one third

of a standard deviation per school year.

Accordingly,

for example, the loss of one third of a school year of

learning would correspond to about 11% of a standard

deviation of lost test results (i.e., 1/3 x 1/3).

In order to understand the economic losses from school

closures, this analysis uses the estimated relationship

between standard deviations in test scores and

individual incomes from a recent international study

(Hampf, Wiederhold and Woessmann, 2017

[8]

This

analysis is based on data from OECD’s Survey of Adult

Skills (PIAAC), the so-called “Adult PISA” conducted by

the OECD between 2011 and 2015, which surveyed

the literacy and numeracy skills of a representative

sample of the population aged 16 to 65. It then

relates labour-market incomes to test scores (and other

factors) across the 32 mostly high-income countries

that participated in the PIAAC survey.

Years of schooling and

earned income

An even more extensive literature examines how

additional years of schooling – which are far easier

to measure than the skills actually acquired – affect

income in the labour market. The strong correlation

between years of schooling and income is probably

one of the most robust ﬁndings of all empirical

economic research. Numerous studies that focus on

identifying the causal effect of additional years of

schooling are quite consistent with simple estimation

of the relationship.

The possible effects of lost school

years are consistent with the literature reported above:

roughly speaking, each school year is associated

with an average of about 10% higher income in many

countries.

| The Economic Impacts of Learning Losses

Table 1•Lost

individual income due to Corona-induced learning loss

Learning loss

(school-year equivalents)

0.25

0.33

0.50

0.67

1.00

Pooled

(0.232)

(0.274)

Lowest

[Greece]

(0.137)

Highest

[Singapore]

(0.501)

1.9%

2.6%

3.9%

5.2%

7.7%

2.3%

3.0%

4.6%

6.1%

9.1%

1.1%

1.5%

2.3%

3.0%

4.6%

4.2%

5.6%

8.4%

1 1%

16.7%

Note:

The values in parentheses in the row headers are the income return per standard deviation of individual test scores.

Source:

Author calculations based on Hampf, Wiederhold and Woessmann, (2017

[8]

), “Skills, Earnings, and Employment: Exploring Causality in the Estimation of

Returns to Skills”,

Large-scale Assessments in Education,

Vol. 5/1, pp. 1-30.

Estimates based on the OECD Survey of Adult Skills

show that income increases by 11.1% per additional

year of schooling in the United States and by 7.5% on

average across the sampled countries (Hanushek et

al., 2015

[9]

: Table 2). If one again considers a

corona-induced loss of one third of a school year,

these results would suggest a loss of income for

the affected pupils of about 2.5-4% over the entire

working life – very similar to the estimates reported

above on the basis of the loss of skills. Moreover,

additional years of schooling are systematically

associated with higher employment and lower

unemployment (see, e.g., (Woessmann, 2016

[10]

)).

These estimates assume that only the cohort currently

in school is affected by the closures and that all

subsequent cohorts resume to normal schooling.

If schools are slow to return to prior levels of

performance, the growth losses will be proportionately

higher.

Slower growth from the loss of skills in today's students

will only be seen in the long term. However, when

viewed over the long term, they assume an enormous

magnitude. In other words, countries will continue to

face reduced economic well-being, even if the schools

immediately return to the pre-pandemic levels of

performance.

These estimates can be applied to individual

countries.

For example, for the United States, if the

student cohorts in school during the 2020 closures

record a corona-induced loss of skills of one tenth of

a standard deviation and if all cohorts thereafter return

to previous levels, the 1.5% loss in future GDP would

be equivalent to a total economic loss of USD 14.2

trillion. By the nature of growth, these losses will not be

felt for some time in the future, but the calculations of

the present value of these growth losses puts them into

current dollar terms.

The overall economic growth effects show that higher

skills of one person do not come at the expense of

the economic opportunities of others. The overall

economic costs of lost learning are not less if they

affect all pupils equally. The notion that lost years of

schooling are not so bad if they affect everyone are

based on the erroneous assumption of a national

economic “cake” of ﬁxed size and that education

largely serves to determine the share of income going

to each individual. But the cake shrinks when everyone

reaches a lower level of education; the entire economy

suffers, not least because of higher burdens on social

security systems and lost tax revenues for social tasks.

Skills and economic growth

Better educational achievement is reﬂected not

only in higher individual incomes but also in higher

national incomes overall. Basic cognitive skills, as

measured in international comparative tests for pupils

in math and science, are probably the most important

long-term determinant of economic growth and thus of

the long-term prosperity of a society (Hanushek and

Woessmann, 2008

[7]

, 2012

[11]

, 2015

[12]

, 2016

[13]

These results on the relationship between educational

performance and economic growth can be used to

calculate projections of the economic costs of learning

losses.

Table 2 uses existing estimates of how skills of the

labour force relate to economic growth to evaluate

the potential aggregate losses of school closures. (The

details of these projections are found in Annex B.) These

estimates come from comparing the GDP expected

across the remainder of the 21st Century with the

given learning losses versus that without such losses.

A learning loss equivalent to one-third of a year of

schooling for the current student cohort is estimated

according to historical growth relationships to mean

1.5% lower GDP on average for the remainder of

the century. The present value of the total cost would

amount to 69% of current GDP for the typical country.

The Economic Impacts of Learning Losses |

Table 2•Long-run

loss in GDP due to Corona-induced learning loss

Learning loss

(school-year equivalents)

0.25

0.33

0.50

0.67

1.00

In % of discounted

future GDP

1.1%

1.5%

2.2%

2.9%

4.3%

In % of current GDP

52%

69%

103%

136%

202%

GDP decrease in year 2100

1.9%

2.6%

3.8%

5.1%

7.5%

Note:

See Annex B for projection methodology.

Source:

Author calculations based on OECD, Hanushek and Woessmann (2015

[14]

Universal Basic Skills: What Countries Stand to Gain.

Distributional issues

Because of the nature of school closures, students

returned to their homes for the duration. This entirely

unexpected event left most schools unprepared for

continued learning. Schools pursued a wide variety

of strategies involving use of technology, provision

of written materials, and hybrid approaches. But

the schooling-at-home approach clearly relies

considerably on the instructional skills of parents, while

the use of technological solutions to ameliorate the

effects of closures depended on broad availability of

tablets, computers, and internet access.

The negative impact of this situation was undoubtedly

greater for students from disadvantaged households.

Low-achieving students will ﬁnd it particularly hard

to acquire new instructional material on their own at

home, without the explanations and support of trained

teachers. It is not simply a matter of closing the “digital

divide” across households. Thus, the prior estimates of

the earnings loss, which apply to students of average

achievement in the average household, are likely to

underestimate the career earnings losses to students

from disadvantaged households and for low-achieving

students. These considerations also hold into the future

to the extent that disruptions in schooling with different

re-opening strategies imply continued pressure on

learning at home.

Note that this differential impact across students does

not necessarily affect the estimates of aggregate losses

in GDP through growth, because those estimates

relate to the average cognitive skills of the population.

The distribution of the rewards from growth would,

nonetheless, likely be skewed by the differential

learning losses across affected families.

Aggregate losses in GDP across G20 nations

The magnitude of the long run losses associated with

the disruption in schooling are truly huge. Most of

the public and governmental attention has focused

on short run issues of unemployment and business

closures. As important as these issues are, they tend to

mask the more serious long-run costs. Table 3 provides

estimates for each of the G20 countries of the present

value of GDP lost over the remainder of the century.

These losses are calculated assuming that just the

grade 1-12 students who faced the initial disruption of

schooling in 2020 are affected and that the education

system returns to 2019 levels for all other past and

future students. The economic losses from 1/3 year of

learning range from an estimated economic downturn

of USD 504 billion in South Africa to USD 15.5 trillion

in China. If the disruption turns out to be greater, these

losses grow proportionately.

The magnitude of these losses requires systematic

and sustained actions to improve the educational

opportunities of the current and future students.

| The Economic Impacts of Learning Losses

Table 3•Present

value of lost GDP due to Corona-induced learning loss for G20 nations

GDP 2019

(billions USD)

Argentina

Australia

Brazil

Canada

China

France

Germany

India

Indonesia

Italy

Japan

Republic of Korea

Mexico

Russian Federation

Saudi Arabia

South Africa

Turkey

United Kingdom

United States

990

1 262

3 092

1 843

22 527

3 097

4 474

9 229

3 197

2 557

5 231

2 206

2 519

3 968

1 609

731

2 350

3 121

20 575

-683

-871

-2 134

-1 272

-15 543

-2 137

-3 087

-6 368

-2 206

-1 765

-3 609

-1 522

-1 738

-2 738

-1 1

-504

-1 621

-2 154

-14 197

Impact of Lost Learning

(billions USD)

-1/3 year learning

-2/3 year learning

-1 347

-1 716

-4 205

-2 507

-30 636

-4 212

-6 084

-12 552

-4 347

-3 478

-7 1

-3 000

-3 426

-5 397

-2 189

-994

-3 196

-4 245

-27 982

Note:

GDP for 2019 is in billions of US dollars in 2017 purchasing power parity (PPP) terms from the World Bank. Present value of lost GDP is based on estimated

difference in GDP for 80 years with lower achieving labour force expected from educational losses of one-third or two-thirds years compared to future GDP

without learning loss. Future losses are discounted at 3 percent. See Annex B for estimation of impacts from lower growth.

Source:

Authors calculations;

World Development Indicators database:

(World Bank, n.d.

[15]

)

https://data.worldbank.org/indicator/NY.GDP.MKTP.PP.KD?name_desc=true

(accessed August 21, 2020).

Making up for learning losses

Both the re-opening strategies of countries and the

opportunities for school improvement vary widely

around the world. Because of the indicated substantial

costs of learning losses, the most immediate – and most

obvious – responses to the COVID-19 situation are to

return to schools wherever epidemiologically feasible

and – where it is not – to implement daily online

instruction rather than leaving children on their own.

The one essential backdrop is, however, that the

current cohort of students will be less prepared for

further schooling and ultimately for the labour force

than they would have been without the pandemic.

Thus, for these students the old status quo will not serve

them well. If these students are to be remediated, it

would require improving the schools, not returning

schools to where they were in 2019.

The large

cross-country variation in the productivity of schools

suggests such improvement is possible.

A second element of the new environment is that less

information is available. Again, while highly variable,

with closures many countries effectively suspended

student assessments and normal school accountability.

With the varied conditions for re-opening, and with

concerns about the impacts of past disruptions, in

some countries there is considerable sentiment for

suspending testing and accountability during the

following year(s). Such actions could have serious

repercussions. Schools will have only imperfect

information about the learning losses suffered,

particularly for disadvantaged students and others

The Economic Impacts of Learning Losses |

hardest hit by the past closures. Furthermore, such a

suspension of testing would threaten important uses of

assessments – uses that have previously been shown

to have positive effects on student learning (Bergbauer,

Hanushek and Woessmann, 2019

[16]

The special nature of the COVID-19 pandemic does

elevate two possible approaches to improving the

schools and to dealing with the deﬁcits of the current

students.

Research from both the developed world and the

developing world has consistently pointed to large

differences among teachers in their effectiveness in

the classroom (Hanushek and Rivkin, 2012

[17]

; Burgess,

2019

[18]

; Harbison and Hanushek, 1992

[19]

; Hanushek,

Piopiunik and Wiederhold, 2019

[20]

). It seems very

likely that this carries over to new modes of instruction.

In particular, while it has yet to be analysed rigorously,

some teachers are undoubtedly better than others

at providing video-based instruction, while others

are more effective at providing in-person instruction.

Policies that recognise differences in effectiveness and

that use more effective teachers in a better manner

would improve overall school performance. For

example, the more effective teachers for video-based

instruction might be given expanded groups of students

(with suitable offsets for the expanded workload of

doing this).

A second change from normal operations of schools

relates to the larger variations in student preparation

that are likely to appear in the majority of classrooms.

As indicated, the closure period, if not the initial

re-opening period, impose larger education burdens

on some students than others – leading to more

variation in the performance levels of students in

each class. While teachers have long recognised

the variation in their incoming students, the substantial

suspension of standard teaching will make this larger.

A natural response (which could be institutionalised) is

to move toward more individualised instruction. Some

countries have already moved closer to a mastery

learning concept. Students would work on speciﬁc

learning modules until they could demonstrate that

they have completely mastered them. At that time, they

would move forward, regardless of what other students

in their classes were doing. Students in the same

classroom could have differentiated learning goals,

ranging from the understanding of basic concepts to

the mastery of deep academic challenges in each

area. Such individualised instruction could be greatly

helped by digital learning technologies that adapt

learning goals to the individuals’ current achievement

levels.

These two examples – while drawing on special

features of the current COVID-19 school

situation – largely reﬂect ideas that have been

discussed over a considerable period. The current

reactions to the pandemic do, however, open the

possibility of moving in directions that improve school

quality and thus offer hope of eliminating the learning

gap faced by today’s students.

Conclusions

As a result of the schools being closed due to the

COVID-19 pandemic, classes were almost universally

disrupted for months in the ﬁrst half of 2020. As pupils

gradually return to school, the high costs of not learning

should be taken into account. The future impact of past

and future learning losses need to be considered when

it comes to the design of mixed in-person and home

learning and when classes are potentially cancelled

again locally or regionally due to newly occurring

infections.

Roughly speaking, research in the economics of

education shows that each additional year of

schooling increases life income by an average of

7.5-10%. In other words, a loss of one third of a school

| The Economic Impacts of Learning Losses

year’s worth of learning would reduce the subsequent

earned income of the pupils concerned by about

3%. Beyond crudely measured school attainment, the

loss in cognitive skills resulting from school closures and

the untested ways of re-opening is the larger issue.

The different ways of estimating the economic costs of

the pandemic for current students provide consistent

estimates of today’s learning challenges.

The costs of school closure and the associated

learning losses go beyond the lower incomes that this

cohort of students can expect. A less skilled work force

also implies lower rates of national economic growth.

A loss of one-third of a year in effective learning for

just the students affected by the closures of early 2020

will, by historical data, lower a country’s GDP by an

average of 1.5% over the remainder of the century.

If the re-opened schools (which also involve new

students) are not up to the same standard as before the

pandemic, the impacts on future economic well-being

will be proportionately larger.

In addition to the economic effects of the cognitive skill

losses emphasised here, there are other potentially

important costs due to losses in social-emotional

development of children, although neither the

magnitude nor the economic impact of these are

currently known.

Immediate concrete measures need to be taken to

provide effective learning for all age groups, albeit

in an adapted format – from improving distance

learning to developing constructive ways to re-open

schools to all children and adolescents. Because

school attendance will likely remain disrupted for

some time to come, the serious costs of not learning

must be considered and comprehensive measures

must be taken to ensure that learning takes place

everywhere again. Indeed, as described, it is possible

and important to build upon the new organisation of

schools to ensure that the schools are actually superior

to the pre-COVID schools.

Unless schools get better, the current students will

be signiﬁcantly harmed. Moreover, the harm will

disproportionately fall on disadvantaged students.

Substantial learning differences across countries,

closely related to institutional structures of their school

systems, indicate that improvements are possible

(Hanushek and Woessmann, 2011

[12]

; Woessmann,

2016

[11]

). Therefore, permanent learning losses are not

inevitable if countries improve the learning gains of

their students in the future.

There is considerable anecdotal evidence that children

from disadvantaged backgrounds and pupils with

learning difﬁculties have a particularly difﬁcult time

coping with the home-learning phase. Due to the

very different pressures, school closures threaten to

become a major burden on the equality of educational

opportunities and lead to increased inequality in

society.

The Economic Impacts of Learning Losses |

Annex A.

Direct evidence on the effects of closed schools

The studies cited previously deal mainly with the

economic effects of skills and years of schooling in

general. In the case of school closures lasting several

months, as in the current case of the COVID-19

pandemic, the question arises whether the learning

that has been missed cannot be made up for. Do the

learning losses due to school closures really have

long-term effects? Analyses of three examples of long

school interruptions – strike-induced school closures,

the German “short school years” of the 1960s, and

long summer holidays – show that this is indeed the

case.

The experience of the

German “Short School Years”

The experiences of the German short school years

in the 1960s show that even a previously planned

reduction in schooling time leaves traces if it lasts

long. In the post-war period, the school year began in

spring in most of Germany’s federal states. In order to

standardise the date of the start of the school year to

the fall nationwide, two short school years were held

in 1966/1967 in many states: the ﬁrst lasted from April

to November 1966, the second from December 1966

to July 1967. In the current literature, the effects of these

short school years are analysed together with those

of the extension of compulsory schooling from eight to

nine years implemented in many states during the same

period.

Based on the German PIAAC data, it can be seen

that the students affected by the two short school

years have indeed received a total of three quarters

of a year less instruction (Hampf, 2019

[26]

: Table 3).

This loss can also be seen in the long-term skills of the

pupils concerned: even in the age group from early

50s to late 60s, the maths skills are still about a quarter

of a standard deviation lower because of the two

years of short schooling (Hampf, 2019

[26]

: Table 4).

In the long term, the short school years have not only

reduced student skills but also their income in the

labour market. The data set “Qualiﬁcations and Career

Progression” shows that the students affected by the

short school years achieved an average of about

5% lower earned income during their working lives

(Cygan-Rehm, 2018

[27]

: Table 4).

In this case, too, it is

therefore true that the loss of schooling has clearly had

negative long-term effects.

Long-term effects of

strike-related school closures

Fortunately, in the past there have not been many

cases of long-term, nationwide school closures. But

there are a number of cases where teacher strikes have

led to school closures lasting weeks or even months.

In several cases, their effects have been studied in

scientiﬁc detail.

In 1990, for example, teachers in the Walloon part

of Belgium went on strike for several months, closing

almost all the schools repeatedly for up to six weeks

at a time over several months. Belot and Webbink

(2010

[23]

) compare the development of the affected

pupils with those in the Flemish part of Belgium, which

was not affected by the strike-related school closures.

Results suggest that the school closures have led to

an increase in grade repetition and, in the long run,

to lower educational attainment, including lower

completion of degrees at higher education levels.

For the Canadian province of Ontario, Baker (2013

[24]

)

shows that teacher strikes have led to signiﬁcantly

lower skill gains of the affected students. Jaume and

Willén (2019

[25]

) looks at particularly long-term effects

of strike-related school closures for Argentina: they

ﬁnd that pupils who were affected by teacher strikes

in primary school later suffer salary losses of 2-3% on

the labour market. They are also more often exposed

to unemployment and work in occupations with lower

skill requirements. Closed schools can therefore indeed

have very long-term negative consequences for the

children and adolescents concerned.

Loss of skills during long

summer holidays

Finally, further insights into the negative effects of

closed schools come from studies of skill development

during summer in countries such as the United States

and Canada that have long summer holidays of two

to three months. To this end, an entire literature has

collected information on knowledge levels of students

at both the beginning and end of the summer holidays.

| The Economic Impacts of Learning Losses

The results show the great importance of closed

schools for the skill development of children and

adolescents, which is usually referred to as summer

learning loss or summer setback. They also reveal

strong differences in summer learning loss between

children from different socio-economic backgrounds

and between pupils with learning difﬁculties and pupils

with strong learning abilities.

On average, over the summer months students suffer

skill losses in the order of about 10% of a standard

deviation.

Closed schools therefore mean not

only stagnation, but a sharp decline (Kuhfeld et al.,

2020

[4]

). This loss of skills is particularly pronounced in

maths, though in reading students from disadvantaged

backgrounds also suffer a pronounced loss of skills. In

contrast, the reading skills of pupils from

socio-economically better-off backgrounds actually

increase slightly over the summer holidays. These

differences in skill loss during the summer holidays

are responsible for a considerable proportion of the

marked socio-economic differences in performance

that arise during school life.

Summary of closure

experiences

Overall, the experience of various cases of continued

school closures – whether due to strikes, short school

years, or long summer holidays – shows that the lack

of schooling has a negative impact on the

long-term opportunities of the children and adolescents

concerned. The experience of the long summer

holidays in particular also suggests that school closures

have widened the gap in skill development. As a result,

there is a great danger that school closures will further

increase future inequality in society.

The Economic Impacts of Learning Losses |

Annex B.

Projection of costs from lower economic growth

The projections of the impacts of learning losses

on growth rely on a simple description of how

skills enter the labour market and have an effect

on the economy.

A range of estimates covering

plausible amounts of learning losses is considered.

These projections parallel projections for gains

in the economies of different countries (OECD,

Hanushek and Woessmann, 2015

[14]

), except here the

consideration is how an economy is affected by a less

skilful workforce.

Learning losses are portrayed as lower cognitive skills

for the cohort of students enrolled in grades 1-12 in

January 2020. The learning losses are presented in

“school-year equivalents” using the rule of thumb that

three years of schooling is equivalent to one standard

deviation of test scores. The projections cover the

range of learning losses from one-quarter to a full year.

The labour force itself will become less skilled than

that in 2019 as increasing numbers of new, poorer

trained people enter the labour market and replace

the more skilled who retire. The estimates assume that

no students before or after those affected by the 2020

closures have lowered skills – i.e., that students outside

of the immediate closure group have a constant

achievement level equal to that of the 2019 workers.

The affected twelve years of students are assumed

to enter the labour force one year at a time starting

in 2021. A worker is assumed to remain in the labour

force for forty years, implying that the labour force

quality falls over the ﬁrst dozen years and will not fully

return to the 2019 quality level until 52 years have

passed (12 years of entry of less skilled students and

40 years until all of the less-skilled workers retire).

The annual growth of the economy is assumed to be

1.98% higher per standard deviation in educational

achievement for the labour force; see (Hanushek

and Woessmann, 2015a

[12]

). This assumes that future

growth follows the pattern of growth of nations

between 1960-2000. Each year into the future,

annual growth is based on the average skill of workers

(which initially changes as new, less skilled workers

enter and later changes when they retire). The estimate

of the loss in GDP with a less skilled workforce

compared to GDP with the existing workforce is

calculated from 2020 until 2100. The growth of the

economy with the current level of skills is projected to

be 1.5%, or the rough average of OECD growth over

the past two decades. The projection is carried out

for 80 years to correspond to the life expectancy of

somebody born in 2020.

Future losses in GDP are discounted to the present

with a 3% discount rate. The resulting present value

of shortfalls in GDP is thus directly comparable to the

current levels of GDP. It is also possible to compare the

gains to the discounted value of projected future GDP

without a pandemic to arrive at the average decrease

in total GDP over the 80 years.

| The Economic Impacts of Learning Losses

Notes

1. For a brief overview of the theoretical foundations of the economic effects of better education and

references to the relevant literature on economics of education since the seminal contributions of

(Schultz, 1961

[28]

; Becker, 1964

[29]

; Mincer, 1974

[52]

), see, for example, (Woessmann, 2016

[10]

(Bradley, and Green, 2020

[57]

) provide an up-to-date overview of research in the economics of

education.

2. For evidence of the economic impact of non-cognitive skills, see for example (Heckman, Stixrud

and Urzua, 2006

[30]

; Lindqvist and Vestman, 2011

[31]

). In addition to the monetary consequences of

education considered here, numerous non-monetary consequences of education have also been

documented (Lochner, 2011

[62]

; Oreopoulos and Salvanes, 2011

[3]

3. Some bounds on overall effects are presented in (Hanushek, 2014

[61]

)

4. Generally, one of two approaches is chosen to address the cognitive skills-income relationship.

On the one hand, some studies measure the skills of students towards the end of high school and

then observe these students again after their transition to the labour market. This way, it is possible

to estimate the association between the skills measured at school age and later income, which

is usually measured in the early years of employment. Examples of this ﬁrst group of studies are

(Murnane, Willett and Levy, 1995

[42]

; Neal and

Johnson

, 1996

[58]

; Mulligan, 1999

[33]

; Murnane

et al., 2000

[32]

; Altonji and Pierret, 2001

[50]

; Chetty et al., 2011

[34]

; Lindqvist and Vestman, 2011

[31]

On the other hand, there are studies that survey the cognitive skills of adults in order to be able

to directly investigate the connection of these skills with current income in the labour market for

all age groups. Examples of the second group of studies are (Leuven, Oosterbeek and Ophem,

2004

[35]

; Hanushek and Zhang, 2009

[36]

, Hanushek and Woessmann, 2012

[11]

; Hanushek et

al., 2015

[8]

; Hanushek, Wiederhold and Woessmann, 2017

[37]

). For overview articles see, for

example, (Bowles, Gintis and Osborne, 2001

[38]

; Hanushek and Woessmann, 2011

[21]

, 2008

[7]

;

Hampf, Wiederhold and Woessmann, 2017

[8]

). Overall, studies of the two approaches come to

very similar conclusions. However, research has shown that the income effects are substantially

underestimated if only persons in the early years after entering the labor market – approximately

up to the age of 35 – are considered (Hanushek et al., 2015

[9]

)

5. The approaches include simulation of achievement models (e.g., (Azevedo et al., 2020

[39]

; Dorn et

al., 2020

[40]

)), extending observations of learning loss over the summer breaks (e.g., (Kuhfeld et al.,

2020

[4]

)) and potentially applying information about prior breaks in schooling because of strikes,

institutional changes, and the like. These last approaches are summarised in the Annex A.

6. A one standard deviation difference in scores would correspond to the difference between the test

score of somebody at the test mean (the 50th percentile) and somebody at the 84th percentile.

One half standard deviation corresponds to the difference in scores of somebody at the 50th

percentile and somebody at the 69th percentile of the test distribution.

7. Note, however, that this correspondence has not been extensively researched and is likely to vary

by grade level, position in the test distribution, and other factors.

8. Hampf, Wiederhold and Woessmann (2017

[8]

) extend and reﬁne the prior estimates in Hanushek

et al. (2015

[9]

). The updated study focuses on causal interpretations of the underlying statistical

models and on the importance of measurement errors in test scores.

The Economic Impacts of Learning Losses |

9. These estimates refer to the relationship that is corrected for measurement errors; estimates dealing

with issues of reverse causality are much larger, while those considering omitted variables are

slightly smaller (Hampf, Wiederhold and Woessmann, 2017

[8]

10. The literature on the returns to years of schooling is so extensive that numerous survey articles have

already dealt with it; see for example (Card, 1999

[59]

; Card, 2001

[41]

; Harmon, Oosterbeek and

Walker, 2003

[42]

; Heckman, Lochner and Todd, 2006

[60]

; Psacharopoulos and Patrinos, 2018

[43]

;

Gunderson and Oreopoulos, 2020

[57]

11. Estimated losses are the present value of income over the century with a discount rate of 3%. For

the general methodology, see (OECD, Hanushek and Woessmann, 2015

[14]

) and the speciﬁcs in

Appendix B.

12. These estimates and much of the discussion of overall costs apply most directly to the more

developed countries where school attendance is close to universal. Developing countries,

particularly with more fragile school systems, will likely face additional challenges.

13. A similar point is made in

https://unesdoc.unesco.org/ark:/48223/pf0000374029.

14. The more recent ﬁndings show that the lack of evidence for long-term effects of the short school

years in an earlier study by Pischke (2007

[44]

) may be due more to methodological issues.

15. For overview articles, see, for example, (Cooper et al., 1996

[45]

; Alexander, Pitcock and Boulay,

2016

[54]

). Important contributions to this literature are, for example (Heyns, 1978

[55]

; Downey, von

Hippel and Broh, 2004

[46]

; Alexander, Entwisle and Olson, 2007

[47]

). A recent contribution is, for

example, (McEachin and Atteberry, 2017

[48]

16. The details of the projection methodology, in somewhat different circumstances, can be found in

(OECD, 2010

[53]

; Hanushek and Woessmann, 2011

[49]

; Hanushek and Woessmann, 2015

[12]

| The Economic Impacts of Learning Losses

Bibliography

Alexander, Karl L., Doris R. Entwisle, Linda S. Olson (2007), “Lasting Consequences of the

Summer Learning Gap”,

American Sociological Review,

Vol. 72, pp. 167-180.

Alexander, Karl, Sarah Pitcock, Matthew Boulay (eds.) (2016),

The Summer Slide: What We

Know and Can Do About Summer Learning Loss,

Teachers College Press, New York, NY.

Altonji, Joseph G., Charles R. Pierret (2001), “Employer Learning and Statistical Discrimination”,

Quarterly Journal of Economics,

Vol. 116/1, pp. 313-350.

Azevedo, João Pedro, Amer Hasan, Diana Goldemberg, Syedah Aroob Iqbal, Koen Geven

(2020), “Simulating the Potential Impacts of Covid-19 School Closures on Schooling and

Learning Outcomes”,

Policy Research Working Paper 9284,

World Bank Group (June),

Washington, DC.

Baker, Michael (2013), “Industrial Actions in Schools: Strikes and Student Achievement”,

Canadian Journal of Economics,

Vol. 46/3, pp. 1014-1036.

Becker, Gary S. (1964),

Human Capital: A Theoretical and Empirical Analysis, with Special

Reference to Education.,

National Bureau of Economic Research, New York, NY.

Belot, Michèle, Dinand Webbink (2010), “Do Teacher Strikes Harm Educational Attainment of

Students?”,

Labour,

Vol. 24/4, pp. 391-406.

Bergbauer, Annika B., Eric A. Hanushek, Ludger Woessmann (2019), “Testing”,

NBER Working

Paper No. 24836,

National Bureau of Economic Research (revised November 2019),

Cambridge, MA.

Bowles, Samuel, Herbert Gintis, Melissa Osborne (2001), “The Determinants of Earnings:

A Behavioral Approach.”,

Journal of Economic Literature,

Vol. 39/4, pp. 1137-1176.

Bradley, Steve, Colin Green (eds.) (2020),

The Economics of Education: A Comprehensive

Overview. 2nd edition,

Academic Press, London.

Burgess, Simon (2019), “Understanding Teacher Effectiveness to Raise Pupil Attainment: Teacher

Effectiveness Has a Dramatic Effect on Student Outcomes—How Can It Be Increased?”,

IZA World of Labor,

Vol. 465.

Card, David (1999), The Causal Effect of Education on Earnings, in Orley Ashenfelter, David

Card (eds.),

Handbook of Labor Economics,

Vol. 3a, North-Holland, Amsterdam,

pp. 1801-1863.

[54]

[47]

[50]

[39]

[24]

[29]

[23]

[16]

[38]

[55]

[18]

[59]

Card, David (2001), “Estimating the Return to Schooling: Progress on Some Persistent Econometric

[41]

Problems”,

Econometrica,

Vol. 69/5, pp. 1127-1160.

Chetty, Raj, John N. Friedman, Nathaniel Hendren, Michael Stepner, Opportunity Insights Team

[1]

(2020), “How Did Covid-19 and Stabilization Policies Affect Spending and Employment? A New

Real-Time Economic Tracker Based on Private Sector Data”,

https://opportunityinsights.org/wp-content/uploads/2020/05/tracker_paper.pdf

(accessed on

22 July 2020).

The Economic Impacts of Learning Losses |

Chetty, Raj, John N. Friedman, Nathaniel Hilger, Emmanuel Saez, Diane Whitmore

Schanzenbach, Danny Yagan (2011), “How Does Your Kindergarten Classroom Affect Your

Earnings? Evidence from Project Star”,

Quarterly Journal of Economics,

Vol. 264/4,

pp. 1593-1660.

[34]

[45]

Cooper, Harris, Barbara Nye, Kelly Charlton, James Lindsay, Scott Greathouse (1996), “The

Effects of Summer Vacation on Achievement Test Scores: A Narrative and Meta-Analytic Review”,

Review of Educational Research,

Vol. 66/3, pp. 227-268.

Cygan-Rehm, Kamila (2018), “Is Additional Schooling Worthless? Revising the Zero Returns to

Compulsory Schooling in Germany”,

CESifo Working Paper 7191,

CESifo, Munich.

Dorn, Emma, Bryan Hancock, Jimmy Sarakatsannis, Ellen Viruleg (2020), Covid-19

and Student

Learning in the United States: The Hurt Could Last a Lifetime,

McKinsey and Co. (June).

Downey, Douglas B., Paul T. von Hippel, Beckett A. Broh (2004), “Are Schools the Great

Equalizer? Cognitive Inequality During the Summer Months and the School Year”,

American

Sociological Review,

Vol. 69/5, pp. 613-635.

Gunderson, Morley, Philip Oreopoulos (2020), "Returns to Education in Developed Countries",

Steve Bradley, Colin Green (eds.),

The Economics of Education: A Comprehensive Overview,

Academic Press, London, pp. 39-51.

Hampf, Franziska (2019),

The Effect of Compulsory Schooling on Skills: Evidence from a Reform

in Germany,

ifo Working Paper 313, ifo Institute for Economic Research, Munich.

Hampf, Franziska, Simon Wiederhold, Ludger Woessmann (2017), “Skills, Earnings, and

Employment: Exploring Causality in the Estimation of Returns to Skills”,

Large-scale Assessments in

Education,

Vol. 5/12, pp. 1-30.

[27]

[40]

[46]

[57]

[26]

[8]

[61]

Hanushek, Eric A. (2014), "The Economic Relevance of Early Childhood Development",

in Jörg F. Maas, Simone C. Ehmig, Carolin Seelmann (ed.),

Prepare for Life! Raising Awareness for

Early Literacy Education,

Stiftung Lesen, Mainz, DE.

Hanushek, Eric A., Marc Piopiunik, Simon Wiederhold (2019), “The Value of Smarter Teachers:

International Evidence on Teacher Cognitive Skills and Student Performance”,

Journal of Human

Resources,

Vol. 54/4, pp. 857-899.

Hanushek, Eric A., Steven G. Rivkin (2012), “The Distribution of Teacher Quality and Implications

for Policy”,

Annual Review of Economics,

Vol. 4, pp. 131-157.

Hanushek, Eric A., Guido Schwerdt, Simon Wiederhold, Ludger Woessmann (2015), “Returns to

Skills around the World: Evidence from PIAAC”,

European Economic Review,

Vol. 73,

pp. 103-130.

Hanushek, Eric A., Guido Schwerdt, Simon Wiederhold, Ludger Woessmann (2017), “Coping

with Change: International Differences in the Returns to Skills”,

Economics Letters,

Vol. 153,

pp. 15-19.

Hanushek, Eric A., Ludger Woessmann (2008), “The Role of Cognitive Skills in Economic

Development”,

Journal of Economic Literature,

Vol. 46/3, pp. 607-668.

Hanushek, Eric A., Ludger Woessmann (2011a), “The Economics of International Differences in

Educational Achievement”, in Eric A. Hanushek, Stephen Machin, Ludger Woessmann (eds.),

Handbook of the Economics of Education,

Vol. 3, North Holland, Amsterdam, pp. 89-200.

[20]

[17]

[9]

[37]

[7]

[21]

Hanushek, Eric A., Ludger Woessmann (2011), “How Much Do Educational Outcomes Matter in

[49]

OECD Countries?”,

Economic Policy,

Vol. 26/67, pp. 427-491.

| The Economic Impacts of Learning Losses

Hanushek, Eric A., Ludger Woessmann (2012), “Do Better Schools Lead to More Growth?

Cognitive Skills, Economic Outcomes, and Causation”,

Journal of Economic Growth,

Vol. 17/4,

pp. 267-321.

Hanushek, Eric A., Ludger Woessmann (2015a),

The Knowledge Capital of Nations: Education

and the Economics of Growth,

MIT Press, Cambridge, MA.

[12]

Hanushek, Eric A., Ludger Woessmann (2016), “Knowledge Capital, Growth, and the East Asian

[13]

Miracle”,

Science,

Vol. 351/6271, pp. 344-345.

Hanushek, Eric A., Lei Zhang (2009), “Quality-Consistent Estimates of International Schooling

and Skill Gradients”,

Journal of Human Capital,

Vol. 3/2, pp. 107-143.

[36]

Harbison, Ralph W., Eric A. Hanushek (1992),

Educational Performance of the Poor: Lessons from

[19]

Rural Northeast Brazil,

Oxford University Press, New York.

Harmon, Colm, Hessel Oosterbeek, Ian Walker (2003), “The Returns to Education:

Microeconomics”,

Journal of Economic Surveys,

Vol. 17/2, pp. 115-155.

Heckman, James J. (2006), “Skill Formation and the Economics of Investing in Disadvantaged

Children”,

Science,

Vol. 312/5782, pp. 1900-1902.

[42]

[6]

[60]

Heckman, James J., Lance J. Lochner, Petra E. Todd (2006), “The Earnings Functions, Rates of

Return and Treatment Effects: The Mincer Equation and Beyond.”, in Eric A. Hanushek, Finis Welch

(ed.),

Handbook of the Economics of Education,

Vol. 1, North Holland, Amsterdam, pp. 307-458.

Heckman, James J., Jora Stixrud, Sergio Urzua (2006), “The Effects of Cognitive and

Noncognitive Abilities on Labor Market Outcomes and Social Behavior",

Journal of Labor

Economics,

Vol. 24/3, pp. 411-482.

Heyns, Barbara (1978),

Summer Learning and the Effects of Schooling,

Academic Press,

New York.

Jaume, David, Alexander Willén (2019), “The Long-Run Effects of Teacher Strikes: Evidence from

Argentina”,

Journal of Labor Economics,

Vol. 37/4, pp. 1097-1139.

Kuhfeld, Megan, James Soland, Beth Tarasawa, Angela Johnson, Erik Ruzek, Jing Liu (2020),

“Projecting the Potential Impacts of Covid-19 School Closures on Academic Achievement”,

EdWorkingPapers No. 20-226,

Annenberg (May), Brown University.

Leuven, Edwin, Hessel Oosterbeek, Hans van Ophem (2004), “Explaining International

Differences in Male Skill Wage Differentials by Differences in Demand and Supply of Skills”,

Economic Journal,

Vol. 114/495, pp. 466-48.

Lindqvist, Erik, Roine Vestman (2011), “The Labor Market Returns to Cognitive and Noncognitive

Ability: Evidence from the Swedish Enlistment”,

American Economic Journal: Applied Economics,

Vol. 3/1, pp. 101-128.

Lochner, Lance (2011), “Nonproduction Beneﬁts of Education: Crime, Health, and Good

Citizenship”, in Eric A. Hanushek, Stephen Machin, Ludger Woessmann (ed.),

Handbook of the

Economics of Education,

Vol. 4, North Holland, Amsterdam, pp. 183-282.

[30]

[55]

[25]

[4]

[35]

[31]

[62]

McEachin, Andrew, Allison Atteberry (2017), “The Impact of Summer Learning Loss on Measures

[48]

of School Performance”,

Education Finance and Policy,

Vol. 12/4, pp. 468-491.

Mincer, Jacob (1974),

Schooling, Experience, and Earnings,

NBER, New York.

[51]

Mulligan, Casey B. (1999), “Galton Versus the Human Capital Approach to Inheritance”,

Journal

[33]

of Political Economy,

Vol. 107/6, pt. 2, pp. S184-S224.

The Economic Impacts of Learning Losses |

Murnane, Richard J., John B. Willett, Yves Duhaldeborde, John H. Tyler (2000), “How Important

Are the Cognitive Skills of Teenagers in Predicting Subsequent Earnings?”,

Journal of Policy

Analysis and Management,

Vol. 19/4, pp. 547-568.

Murnane, Richard J., John B. Willett, Frank Levy (1995), “The Growing Importance of Cognitive

Skills in Wage Determination”,

Review of Economics and Statistics,

Vol. 77/2, pp. 251-266.

Neal, Derek, William R. Johnson (1996), “The Role of Pre-Market Factors in Black-White

Differences”,

Journal of Political Economy,

Vol. 104/5, pp. 869-895.

[32]

[52]

[58]

[5]

OECD/Global Education Innovation Initiative, Harvard University (2020),

Global Education

Innovation Initiative at Harvard and OECD Rapid Assessment of COVID-19 Education Response.

OECD/Hanushek, Eric A., Ludger Woessmann (2010),

The High Cost of Low Educational

Performance: The Long-run Economic Impact of Improving PISA Outcomes,

PISA, OECD

Publishing, Paris,

https://doi.org/10.1787/9789264077485-en.

OECD/Hanushek, Eric A., Ludger Woessmann (2015),

Universal Basic Skills: What Countries

Stand to Gain,

OECD Publishing, Paris,

https://doi.org/10.1787/9789264234833-en.

Oreopoulos, Philip, Kjell G. Salvanes (2011), “Priceless: The Nonpecuniary Beneﬁts of

Schooling.”,

Journal of Economic Perspectives,

Vol. 25/1, pp. 159-184.

Pischke, Jörn-Steffen (2007), “The Impact of Length of the School Year on Student Performance

and Earnings: Evidence from the German Short School Years”,

Economic Journal,

Vol. 117/523,

pp. 1216-1242.

Psacharopoulos, George, Harry Anthony Patrinos (2018), “Returns to Investment in Education:

A Decennial Review of the Global Literature”,

Education Economics,

Vol. 26/5, pp. 445-458.

Schultz, Theodore W. (1961), “Investment in Human Capital”,

American Economic Review,

Vol. 51/1, pp. 1-17.

Woessmann, Ludger (2016), “The Economic Case for Education”,

Education Economics,

Vol. 24/1, pp. 3-32.

Woessmann, Ludger (2016), “The Importance of School Systems: Evidence from International

Differences in Student Achievemen”,

Journal of Economic Perspectives,

Vol. 30/3, pp. 3-32.

Woessmann, Ludger, Vera Freundl, Elisabeth Grewenig, Philipp Lergetporer, Katharina Werner,

Larissa Zierow (2020), “Bildung in der Coronakrise: Wie haben die Schulkinder die Zeit der

Schulschließungen verbracht, und welche Bildungsmaßnahmen befürworten die Deutschen?

[Education in the corona crisis: How did the schoolchildren spend the time the schools were

closed and which educational measures do the Germans advocate?]”,

ifo Schnelldienst,

Vol. 73/9.

World Bank (n.d.),

World Development Indicators database,

World Bank,

https://data.worldbank.org/indicator/NY.GDP.MKTP.PP.KD?name_desc=true

(accessed on 21 August 2020).

[56]

[14]

[3]

[44]

[43]

[28]

[10]

[22]

[2]

[15]

| The Economic Impacts of Learning Losses

This brochure was original published as the following OECD Education Working Paper: Hanushek E and L. Woessman, (2020), “The economic impacts of

learning losses”,

Education Working Papers,

No. 225, OECD Publishing, Paris,

https://doi.org/10.1787/21908d74-e

This paper should not be reported as representing the ofﬁcial views of the OECD or of its member countries. The opinions expressed and arguments

employed herein are those of the author(s).

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.

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

This paper has been authorised by Andreas Schleicher, Director of the Directorate for Education and Skills, OECD.

Photo credits: © Shutterstock/Phuttharak; © Shutterstock/Halfpoint; © Shutterstock/saravutpics; © Shutterstock/Zorro Stock Images

The OECD is compiling data, information, analysis and recommendations regarding the health, economic, ﬁnancial and societal challenges posed by the

impact of coronavirus (COVID-19). Please visit our

dedicated page

for a full suite of coronavirus-related information.

The Economic Impacts of Learning Losses |

For more information, contact

[email protected]

Eric A. Hanushek,

Hoover Institution, Stanford University;

[email protected]

Ludger Woessmann,

ifo Institute and University of Munich;

[email protected]

Connect with us:



[email protected]



https://oecdedutoday.com/



https://www.oecd-ilibrary.org/education



@OECDEduSkills



OECD Education and skills



@oecd_education_skills

Visit