How should we measure the development of human capital in children?

The academic, social and emotional skills that children acquire during their upbringing play a critical role in their later lives. For most youngsters, the pandemic has disrupted both school and home life, reigniting debate about how best to quantify these elements of ‘human capital’.

A child’s ‘human capital’ is a collection of their skills and abilities. Some children are stronger academically; some have better social and emotional skills; and others excel at practical subjects. No single measure can capture all of these things – the best measure to use depends on what we’re interested in.

What is human capital and why does it matter?

When economists talk about children’s human capital, they are really referring to their collection of skills and abilities: the attributes that help children (and people in general) to accomplish their goals in life.

Human capital includes concepts as varied as a child’s intelligence and academic ability, their capacity to make friends or concentrate in class, or their physical and mental health. While there is a huge range of skills and abilities that fall under this umbrella, many researchers group them into three broad categories of development, namely cognition; social and emotional development; and health.

The Covid-19 crisis has underlined the importance of human capital development. Months of home learning have left many children, especially in disadvantaged families, academically behind where they would typically be. Across a range of studies in England, researchers have estimated that the first set of school closures cost students roughly one or two months of expected academic progress (Education Endowment Foundation, 2021).

Other children have suffered from the social and emotional consequences of living in lockdown Blanden et al, 2021). And others will have seen their physical health deteriorate during the pandemic, whether due to Covid-19 itself or because of the lack of opportunities to access normal healthcare and physical activities (Hefferon et al, 2021).

Failing to develop these skills and abilities as quickly as usual – or even seeing them deteriorate – can have lifelong consequences. Many studies show that cognitive development in childhood is important for adult outcomes such as earnings and the level of education attained (Almond and Currie, 2011; Heckman et al, 2010; Hanushek and Woessmann, 2012).

More recent evidence points to the importance of emotional development and social skills in determining adult outcomes. Attributes like self-control and emotional regulation are not only important in and of themselves for child wellbeing: they also help to predict employment and earnings later in life (Daniel et al, 2020; Goodman et al, 2015).

Additionally, limited social and emotional skills are a strong predictor of poor adult mental and physical health, as well as crime (Attanasio et al, 2020; Baker et al, 2019). Promoting health in childhood has benefits for educational attainment, earnings and life expectancy (Bhalotra et al, forthcoming).

It is tempting to argue that there should be less concern about a loss of skills in younger children, who have longer to catch up. But there is strong evidence that human capital ‘builds on itself’, so a loss of skills early on can have outsized impacts on the level of human capital a child is eventually able to attain (Cunha et al, 2010). For example, a child who misses out on foundational skills like reading or listening to the teacher will find it more difficult to make the most of lessons for the rest of their schooling.

How do we measure different parts of children’s human capital?

Unlike many other outcomes, such as earnings or crime, human capital can never be measured directly. Instead of a single unambiguous measurement, researchers observe tests, characteristics and other measures that are influenced by these underlying concepts. A child’s performance on any test or task will reflect a combination of their underlying human capital (often many different aspects of it) and other circumstances affecting the test.

The role of circumstances can be profound. Imagine taking a test in a quiet library after a good night’s sleep compared with doing the same assessment in the middle of a noisy sports hall with no heating. Even random chance can affect results: researchers in Israel, for example, have found that children taking their exams on a particularly polluted day earn tens of thousands of dollars less over their lifetimes as a result (Ebenstein et al, 2016). Other research has shown that taking SATs (standard tests for American school children) on a particularly hot day can reduce test scores by up to 14% (Park, 2020).

Because of the role of circumstances and chance, the best way of measuring human capital is to combine the results of many different measures and use statistical techniques to work out what they have in common. These common factors can then give researchers a better grasp of the underlying domain of human capital they wish to capture (Cunha et al, 2010).

Within the broad categories of health, cognition and socio-emotional development, there are many different dimensions of human capital. For example, cognition can include concepts from executive functioning skills (like planning and prioritising) to mathematical ability. Social and emotional development includes emotional stability and social skills, alongside abilities like persistence and effort. Health includes chronic conditions (such as deafness or poor eyesight), acute health problems (like illness or disease), and fitness and nutrition.

Further, the different components of human capital can be general abilities that are relevant to much of everyday life, or they can be very specialised to the task at hand. For example, woodworking skills are important for being a carpenter, but matter much less to a nurse.

Even when researchers find a measure of human capital that works well, its usefulness is often limited to specific age groups and cultural contexts. For example, child development psychologists use a number of assessments that can measure cognition and related skills from as young as eight weeks old (these include the ‘WPPSI-IV’, ‘Bayley-III’ and the ‘Ages and Stages Questionnaire’). But the definition of cognition and what is considered ‘smart’ change as children get older. Being good at placing blocks in the right shaped hole is impressive in a toddler, but less so in a teenager.

How well do test scores measure human capital?

All of these difficulties mean that there is no one perfect tool for measuring human capital. Instead, researchers – and society as a whole – try to choose measures that are appropriate to the task at hand. These choices involve trade-offs between the reliability of a test at picking up on the dimensions of human capital it is aimed at; how close those parts of human capital are to what the researcher actually wants to measure; and how difficult it is to administer the assessment.

One of the most influential sets of measures of human capital in our society is the test scores and grades that children and young people receive in school. These have substantial implications for children and young people. One study found that just missing out on a C grade at GCSE English Language meant that students were much more likely to drop out of education without employment three years later compared with their peers who scored just above them (Machin et al, 2020).

This example illustrates the challenge of achieving perfectly fair outcomes using high-stakes tests. Scores can be influenced by random chance – air pollution, weather, illness – on the day of the assessment. This means that a child who misses the C grade by one point may be, in reality, no different to one who just made the mark. Their very different outcomes later in life illustrate how much weight is placed on these grades.

Measures of human capital reflect context as well as skill

There is another set of challenges around how far to take context into account. One example of this is a child’s age: just as the age of a child or teenager affects how impressive their block-to-hole matching performance is, age affects test scores in the schooling years.

Children who are born during summer months, on average, perform worse than their autumn-born peers, who are nearly a full year older at the time they sit the tests (Crawford et al, 2014). The way that tests are designed also matters; men on average respond better to competitive environments than women, and so high-stakes exams often given an exaggerated measure of the actual gender difference in skills and abilities (Niederle and Vesterlund, 2010).

This argument could be pushed further still. Should a child who is raised in a non-English-speaking family be given extra credit on their reading test? Should a teacher give extra time to pupils who don’t have access to private tutoring? Most topically, should pupils whose education has been most disrupted by Covid-19 be given some sort of leeway in exams?

These are difficult questions, and there is no single right answer. One way to start thinking through the issues is to focus on what types of human capital a test is measuring, and what we want it to measure. If we are using a reading test to assess applicants for a copy-editing job, we are likely to want to capture that very specific dimension of human capital: in this case, there is less scope for incorporating contextualising factors. But if instead we are using the tests to screen children for developmental delay, it makes sense to account for whether the child comes from an English-speaking home.

This highlights the importance of setting assessments appropriately, recognising that performance will always be affected both by context and by skill. In the United States, for example, there is an enduring debate, dating back to the 1980s, about the use of standardised assessments like SATs, where children from disadvantaged, black and Hispanic backgrounds often under-perform compared with their more affluent or white peers. Historically, some of this gap is related to questions that test very specific cultural knowledge, such as which sportsmen participate in a regatta.

But another part of the gap reflects the accumulated educational disadvantages that ethnic minority and lower-income communities face. All else being equal, it’s easier to answer timed logic questions if you’ve had the benefit of good schooling in the years before. In this case, the test is picking up a real difference in human capital, and one that colleges might well care about. But that is not the same as saying that it is free from the influence of inequalities. And colleges may also be interested in a prospective student’s potential, and not just what they already know.

Measurement is never perfect – but it can be essential

Imperfection is not necessarily an argument for scrapping these assessments. It is not possible or practical to measure all of the different dimensions of human capital across the entire population perfectly. We do need ways of assessing pupils’ human capital development and of sorting them into educational destinations. And standardised tests can provide a less socially biased measure of students’ abilities than other, less formal assessments such as unmoderated teacher assessments (Murphy and Wyness, 2019; Lozano, 2020).

But it does mean that we need to keep in mind the limitations of test scores, and to ensure that our society and educational system offer second (and third and fourth) chances to access education and training later in life. Within the context of exam disruption over the last two years, it is important to bear in mind that no academic qualification is a perfect measure of ability. While measurement can help us to understand the scale of the problem facing us, we should not forget that the much more pressing matter is the loss of skills, not a change in test scores.

Where can I find our more?

The economics of early childhood development: IFS public lecture exploring how government policy in the early years can shift human capital development.
The case for investing in disadvantaged young children: Economics Nobel laureateJames Heckman gives an overview of the ‘Heckman framework’ for thinking about human capital development.
Assessing and monitoring pupil progress: Toolkit from the Education Endowment Foundation on how to monitor progress in the schooling years.
Reducing parental conflict measures selector: An interactive toolkit developed by the Early Intervention Foundation that gives an overview of different options for measuring child outcomes.