The Human Capital Perspective: How Employee Knowledge, Skills, and Capabilities Create Economic Value

Article 08 treated employee knowledge as a strategic resource and asked when it gives one firm an advantage over its rivals. This article goes back to an earlier question: how knowledge and skill create economic value at all, and how anyone could tell. The modern answer starts with Theodore W. Schultz's presidential address to the American Economic Association, delivered on 28 December 1960 and published in March 1961 as "Investment in Human Capital."

Schultz began with a measurement problem. "It has been widely observed," he wrote, "that increases in national output have been large compared with the increases of land, man-hours, and physical reproducible capital. Investment in human capital is probably the major explanation for this difference" (Schultz, 1961, p. 1). Calling the unexplained gap "resource productivity," he added, "gives a name to our ignorance but does not dispel it" (p. 6). On the rough estimates then available, he suggested that between 36 and 70 percent of the previously unexplained rise in American labour earnings came from returns to workers' additional education (p. 13).

Two features of the paper tend to disappear in later summaries. One is its breadth. Schultz listed five categories of investment: health facilities and services; on-the-job training, "including old-style apprenticeship organized by firms"; formal education; study programmes for adults outside firms, such as agricultural extension; and the "migration of individuals and families to adjust to changing job opportunities" (p. 9). Health and migration belong on the list for the same reason as schooling: each uses resources now to raise productive capacity later.

The other is cost. The largest cost of learning, Schultz argued, is often the earnings given up while doing it; in the United States "well over half of the costs of higher education consists of income foregone by students" (p. 11). The point carries straight into management. A firm that counts only trainer fees understates what training costs, because the largest item is usually the output employees do not produce while they learn.

Schultz also met the objection that would follow the idea for decades. "Free men are first and foremost the end to be served by economic endeavor," he wrote; "they are not property or marketable assets" (p. 2). The capital belongs to the worker: "Laborers have become capitalists not from a diffusion of the ownership of corporation stocks, as folklore would have it, but from the acquisition of knowledge and skill that have economic value" (p. 3). A firm can pay for human capital but cannot own it, a point that returns when we reach the balance sheet. Schultz, who shared the 1979 Nobel Memorial Prize in Economic Sciences with Sir Arthur Lewis for research on developing countries, came to the subject from the economics of poor agricultural societies, which is partly why his list begins with health.

Schultz supplied the question; Gary Becker supplied a theory of how individuals and firms decide. He defined investing in human capital as "activities that influence future real income through the imbedding of resources in people" (Becker, 1962, p. 9). Article 08 used the distinction between general and firm-specific skills to ask which skills protect an advantage. Becker used it to ask who pays. In a competitive labour market, firms "would provide general training only if they did not have to pay any of the costs," because a trained worker could take the skill elsewhere. Trainees would pay instead, through lower wages while training, so "the cost as well as the return from general training would be borne by trainees, not by firms" (p. 13). The costs and returns of firm-specific training would be shared.

Jacob Mincer turned the theory into the most widely used tool in labour economics. In Schooling, Experience, and Earnings (1974) he showed that, under simplifying assumptions, the logarithm of earnings rises in proportion to years of schooling, with "the rate of return as the coefficient of proportionality" (p. 11). Adding years of experience and its square, to capture on-the-job learning that slows over a career, gives the earnings function estimated in thousands of studies since:

ln(earnings) = a + r × (years of schooling) + b₁ × (experience) − b₂ × (experience)²

Pooling estimates from 139 countries between 1950 and 2014, George Psacharopoulos and Harry Patrinos (2018) put the average private return to an extra year of schooling at about 9 percent (8.8 percent in their tables), with 8.1 percent in South Asia and returns for women about two percentage points higher than for men (pp. 445, 452).

The coefficient r is easy to over-read. Mincer's interpretation assumed that all investment costs are time costs, that working life does not depend on schooling, and that each person's rate of return is fixed (Mincer, 1974, pp. 7-11). James Heckman, Lance Lochner, and Petra Todd (2006) tested such assumptions and rejected them: the coefficient is "a growth rate of market earnings with years of schooling and not an internal rate of return measure, except under stringent conditions." Reading "the return to a degree is 9 percent" as an investment yield takes the same unwarranted step.

The harder problem is causal. People who stay longer in education differ in ability, background, and motivation, and if those traits raise earnings, the coefficient partly measures them. Reviewing studies of identical twins and natural experiments such as changes in compulsory schooling laws, David Card (1999) concluded that the causal return "is not much below" the conventional estimate. Twin studies suggested an upward ability bias "on the order of 10%," and studies exploiting school reforms found returns 20 to 40 percent above the conventional figure (p. 1802). The link between schooling and earnings is therefore not mainly an artefact of who chooses to study. How schooling produces the higher wage is the separate question signaling theory raises.

Wages are the usual evidence for human capital because they are recorded everywhere and productivity is not. That convenience assumes workers are paid what they add. Lorraine Dearden, Howard Reed, and John Van Reenen (2006) tested the assumption on British industries from 1983 to 1996. A one-percentage-point rise in the share of workers trained was associated with about 0.6 percent more value added per worker but only about 0.3 percent higher wages, implying, in their words, that earlier wage-based studies had "underestimated the importance of training." Jozef Konings and Stijn Vanormelingen (2015), using Belgian firm-level data, estimated that a trained worker was 23 percent more productive than an untrained one but earned 12 percent more.

Neither result fits Becker's competitive benchmark, in which workers collect the whole return to general training. Daron Acemoglu and Jörn-Steffen Pischke (1998) offered an explanation. The current employer has "superior information about the worker's ability relative to other firms," so rivals do not bid wages up to the worker's full productivity, and the employer keeps part of the gain, which gives it a reason to pay even for portable skills. Comparing German apprentices who left their training firm voluntarily with those who left for military service, a departure that said nothing about ability, they reported that the predictions of the specific human capital model "are rejected, while our model receives support from the data." The firm's share of the return thus depends on labour-market frictions, and it shrinks where rivals can observe skill cheaply. The calculator below treats that share as an explicit input.

In 1973 Michael Spence published a model in which education raises wages without raising productivity at all. It starts from the employer's side. Because it takes time to discover what a recruit can do, hiring is an investment made under uncertainty; in Spence's phrase, "to hire someone, then, is frequently to purchase a lottery" (Spence, 1973, p. 356). Employers therefore reward observable characteristics that predict productivity, and some of these, which Spence called signals, can be acquired at a cost. The model turns on one condition, set out under the heading "A Critical Assumption": a signal separates applicants only if its costs are negatively correlated with productive capability (p. 358). If abler people find study less costly in effort and time, they acquire more of it, and an employer who pays for education is, correctly, paying for ability.

Two implications follow. In Spence's numerical example education adds nothing to the worker's real marginal product (p. 364), yet acquiring it is privately profitable. And the outcome can be wasteful: once a standard sorts people, raising it improves the sorting "not one bit" while using up resources, so that "systematic overinvestment in education is a distinct possibility" (pp. 364-365, 368). Kenneth Arrow (1973) set out a related screening view the same year, and Spence shared the 2001 Nobel prize with George Akerlof and Joseph Stiglitz for analyses of markets with asymmetric information. The table shows where the two theories part company.

The table summarizes the predictions usually drawn from each theory. It is this article's synthesis, not a table from either literature.

Because the theories share their headline prediction, testing between them is hard, and the evidence points both ways. John Tyler, Richard Murnane, and John Willett (2000) exploited the General Educational Development (GED) credential for US high-school dropouts. States set different passing scores, so people with identical scores held the credential in some states but not in others, which held skill constant. The credential alone raised the earnings of young white dropouts by 10 to 19 percent, with no statistically significant effect for minority dropouts (p. 431). Here, at least, a credential carried value beyond the skills it certified.

Other evidence limits signaling's share. If employers first pay for schooling because ability is hard to see, wages should come to depend less on schooling and more on hard-to-observe correlates of ability as employers learn. Joseph Altonji and Charles Pierret (2001) found that pattern among young US men, and in an earlier paper concluded that "the portion of the return to education that could reflect signaling of ability is limited" (Altonji & Pierret, 1996). At the other pole, Bryan Caplan (2018) argues that education mainly certifies intelligence, work ethic, and conformity, and puts signaling's share of the payoff at 80 percent, a deliberate challenge to the mainstream view, not a consensus figure. Most labour economists attribute the larger part of the schooling premium to skills that schooling builds, while accepting that credentials carry signaling value at particular points. The split is unresolved and probably varies by country, level of education, and occupation.

For a manager, the debate bites more narrowly. Signaling is a problem of hiring strangers: a firm training people it has employed for years already observes their ability, so its training cannot signal anything to the firm itself. Spence's problem reappears in degree requirements that screen for traits a job does not need, and in programme evaluation, where a firm that selects its strongest people for development and then credits the programme with their performance repeats the ability-bias error. The case below makes that risk concrete.

If knowledge and skill account for much of what a service or knowledge-intensive firm is worth, their absence from its balance sheet needs explaining. IAS 38 Intangible Assets, issued by the International Accounting Standards Board, defines an asset as a resource "controlled by an entity as a result of past events" from which "future economic benefits are expected to flow to the entity" (para. 8). Human capital fails on control, as paragraph 15 states:

An entity may have a team of skilled staff and may be able to identify incremental staff skills leading to future economic benefits from training. The entity may also expect that the staff will continue to make their skills available to the entity. However, an entity usually has insufficient control over the expected future economic benefits arising from a team of skilled staff and from training for these items to meet the definition of an intangible asset.

Specific management or technical talent is likewise unlikely to qualify "unless it is protected by legal rights to use it," and expenditure on training activities is recognised as an expense when incurred (para. 69(b)). The standard does not deny that training yields future benefits; it denies that the firm controls them. Even a company that buys another and pays for its people cannot record them: IFRS 3 Business Combinations treats an assembled workforce as not identifiable, so "any value attributed to it is subsumed into goodwill" (para. B37). This is Schultz's ownership point in accounting form, with a reliability concern added: a balance-sheet figure for people would rest on management's own forecasts of productivity and retention. Human resource accounting, developed from the late 1960s by Eric Flamholtz and colleagues, built valuation methods used mainly inside firms (Flamholtz, Bullen, & Hua, 2002).

The cost of this caution falls on decisions: a firm that cuts training reports higher profit that year, while the lost capability appears later, unlabelled. Baruch Lev and Feng Gu, introducing their 2016 book, wrote that value-creating investments in intangibles "are considered regular expenses," and that managers chasing quarterly targets end up "occasionally postponing or cutting advertising, R&D, employee training, or maintenance outlays to 'make the numbers'" (Lev & Gu, 2016a). Their wider claim, that financial reports have lost most of their relevance to investors, is disputed (Radin, 2016). Alex Edmans (2011) offers market evidence that investors underprice what the accounts omit: a value-weighted portfolio of the "100 Best Companies to Work For in America" earned an annual four-factor alpha of 3.5 percent from 1984 to 2009, from which he concluded that "the stock market does not fully value intangibles." The study measures employee satisfaction, not skill, and returns of this kind may fade once documented.

The response so far has been disclosure, not recognition. Since November 2020 companies registered with the US Securities and Exchange Commission have had to describe their "human capital resources," including any measures or objectives management focuses on, where material (SEC, 2020), and ISO 30414:2025 has replaced the 2018 international guidelines for human capital reporting. Both leave the balance sheet unchanged. Human capital reporting is taken up in the governance, ethics, and ESG module later in this series.

Bangladesh shows why the measurement questions matter in practice. In the World Bank's Human Capital Index 2020, a child born in Bangladesh could expect to complete 10.2 years of school by age 18; adjusted for how much students actually learn, measured by harmonized test scores, that fell to the equivalent of 6.0 years. The index value of 0.46 means she would grow up to be 46 percent as productive as she could be with complete education and full health, below the South Asian average (World Bank, 2020). That gap separates what a Mincer equation counts from what Schultz meant by human capital.

The labour market adds a tension. The working-paper annex to Psacharopoulos and Patrinos records a 10.0 percent return to schooling for Bangladesh on 2006 data, yet the Bangladesh Bureau of Statistics' Labour Force Survey 2022 found unemployment highest among the most educated. National unemployment was 3.6 percent (Prothom Alo, 29 March 2023). Among people with tertiary education it was 12 percent, about 800,000 people, up from 11.2 percent in 2016-17, against 8.87 percent for those with higher secondary and 2.82 percent for those with secondary education (The Daily Star, 26 October 2023). Economists quoted in the coverage pointed to a gap between what degree programmes teach and what employers need.

The figures alone cannot decide between the two theories. A human capital reading sees a quality problem: degrees have expanded faster than the skills they certify, which is consistent with the gap between 10.2 years of schooling and 6.0 learning-adjusted years. A signaling reading sees, in part, a queue: where secure formal jobs are scarce and recruitment relies on examinations and degree requirements, people acquire credentials to stay in contention, and some wait unemployed for the jobs those credentials ration. Graduates may also come from households able to support a longer search. These are this article's interpretations, not findings of the survey, but for employers they point the same way: in this labour market a degree is weak evidence of job-ready skill, which raises the value of an employer's own assessment and training.

The calculation treats one trainee's programme as an investment, adjusting the HR head's figure for how much of the gap the programme caused, how much of the gain is paid back in salary, and how long the graduate stays. Every input is a hypothetical assumption for the constructed case, not a benchmark for Bangladeshi banks.

Net present value per trainee = −C + the sum over years t = 1 to 5 of [G × a × (1 − s) × r^t] ÷ (1 + i)^t

With the full gap attributed to the programme, the bank's share of each year's gain is Tk 700,000 × 1 × 0.5 = Tk 350,000, before retention and discounting. Choose an assumption set below to see the result. All figures are rounded to the nearest taka.

The break-even question. A board may find it more useful to ask what share of the gap the programme must cause in order to pay for itself. At 85 percent retention the answer is about 61 percent; at 95 percent, about 46 percent. At 65 percent no share is high enough, because even full attribution leaves the five-year value per trainee below its cost. Retention matters as much as the theoretical dispute.

What the calculation leaves out. It ignores benefits after year five, which biases it against the programme, along with the cost of recruiting lateral hires, effects on colleagues' productivity, and business that former trainees send back. It also treats retention as constant, although departures often cluster when service agreements expire. The most valuable improvement would be evidence on attribution. Comparing graduates with candidates who narrowly missed selection but joined through ordinary recruitment, or varying the programme between cohorts, would turn the CFO's objection into a question the bank can answer.

The points below are this article's synthesis of the evidence above, not findings from a single study.

1. Count the whole cost. Include the output people do not produce while they learn; a programme cheap in fees can be expensive in time.
2. Separate selection from development before claiming a return. Comparing trained high-fliers with everyone else measures the selection process as much as the training. Comparison groups, phased roll-outs, or near-miss candidates give a fairer test.
3. Treat pay and retention as part of the investment case. The firm keeps only the share of gains that wage-setting and turnover leave it.
4. Do not let the accounting treatment set the training budget. Expensing makes training the easiest cost to cut in a weak quarter; reporting a few human capital measures to the board beside the budget makes the trade-off visible.
5. Use credentials with care in hiring. Where degrees mainly ration access to jobs, work samples and structured assessments test the skill directly.

More than sixty years after Schultz's address, the core claim of human capital theory, that knowledge and skill are produced by investment and earn a return, is about as well supported as anything in labour economics. What remains contested is the mechanism behind any particular return and who captures it. The signaling debate has narrowed rather than closed: most evidence puts signaling's share of the education premium well below Caplan's estimate but above zero, and the share varies with the labour market. The accounting position has not moved. Standard setters still treat people as a resource the firm cannot control, which is legally sound and economically awkward.

For the HR function, these unsettled questions lead to one practical demand. The case for investing in people has to be made with evidence that financial statements will never supply: what a programme caused, how long people stay, and how much of the gain the organization keeps. A function that cannot produce that evidence will see its budget judged on the only number the accounts report, which is the expense.

Select a training, graduate, or development programme in an organization you know well or can research in depth. You may disguise the organization's name. Prepare a report of 2,500 to 3,000 words for its board that:

1. describes the programme and estimates its full cost per participant, including output forgone during training, stating how each figure was obtained;
2. estimates the programme's benefits using the structure of this article's calculator or a justified alternative, with every assumption stated and sourced or labelled as an assumption;
3. assesses how much of any observed performance difference the programme caused, addressing selection and signaling explicitly, and proposes a practical design for testing attribution;
4. analyses who captures the return, considering pay, retention, and the labour market for the skills involved;
5. explains how the programme is treated in the financial statements under IAS 38 or the applicable framework, how that treatment affects budget decisions, and which human capital measures the board should receive instead.

Use at least five academic sources, including at least three from this article's reference list. Do not invent data: where information is unavailable, say so, use clearly labelled assumptions, and show how sensitive your conclusion is to them.

Assessment rubric (100 marks)

1. Acemoglu, D., & Pischke, J.-S. (1998). Why do firms train? Theory and evidence. Quarterly Journal of Economics, 113(1), 79–119. https://doi.org/10.1162/003355398555531 (Quotations checked against NBER Working Paper 5605, 1996: https://www.nber.org/papers/w5605)
2. Altonji, J. G., & Pierret, C. R. (2001). Employer learning and statistical discrimination. Quarterly Journal of Economics, 116(1), 313–350. https://doi.org/10.1162/003355301556329. The quotation on signaling is from Altonji, J. G., & Pierret, C. R. (1996). Employer learning and the signaling value of education (NBER Working Paper 5438). https://www.nber.org/papers/w5438
3. Becker, G. S. (1962). Investment in human capital: A theoretical analysis. Journal of Political Economy, 70(5, Part 2), 9–49. https://doi.org/10.1086/258724
4. Card, D. (1999). The causal effect of education on earnings. In O. Ashenfelter & D. Card (Eds.), Handbook of labor economics (Vol. 3A, pp. 1801–1863). Elsevier. https://doi.org/10.1016/S1573-4463(99)03011-4
5. Dearden, L., Reed, H., & Van Reenen, J. (2006). The impact of training on productivity and wages: Evidence from British panel data. Oxford Bulletin of Economics and Statistics, 68(4), 397–421. https://doi.org/10.1111/j.1468-0084.2006.00170.x
6. Konings, J., & Vanormelingen, S. (2015). The impact of training on productivity and wages: Firm-level evidence. Review of Economics and Statistics, 97(2), 485–497. https://doi.org/10.1162/REST_a_00460
7. Mincer, J. (1974). Schooling, experience, and earnings. National Bureau of Economic Research; distributed by Columbia University Press. https://www.nber.org/books-and-chapters/schooling-experience-and-earnings
8. Psacharopoulos, G., & Patrinos, H. A. (2018). Returns to investment in education: A decennial review of the global literature. Education Economics, 26(5), 445–458. https://doi.org/10.1080/09645292.2018.1484426 (Bangladesh estimate from the annex to World Bank Policy Research Working Paper 8402.)
9. Schultz, T. W. (1961). Investment in human capital. American Economic Review, 51(1), 1–17. https://www.jstor.org/stable/1818907
10. Spence, M. (1973). Job market signaling. Quarterly Journal of Economics, 87(3), 355–374. https://doi.org/10.2307/1882010

Other works cited

• Arrow, K. J. (1973). Higher education as a filter. Journal of Public Economics, 2(3), 193–216. https://doi.org/10.1016/0047-2727(73)90013-3
• Caplan, B. (2018). The case against education: Why the education system is a waste of time and money. Princeton University Press. The 80 percent estimate is from Caplan, B. (2018, February 19). Is education worth it? My opening statement for the Caplan-Hanushek debate. EconLog, Econlib. https://www.econlib.org/archives/2018/02/is_education_wo.html
• Edmans, A. (2011). Does the stock market fully value intangibles? Employee satisfaction and equity prices. Journal of Financial Economics, 101(3), 621–640. https://doi.org/10.1016/j.jfineco.2011.03.021
• Flamholtz, E. G., Bullen, M. L., & Hua, W. (2002). Human resource accounting: A historical perspective and future implications. Management Decision, 40(10), 947–954. https://doi.org/10.1108/00251740210452818
• Heckman, J. J., Lochner, L. J., & Todd, P. E. (2006). Earnings functions, rates of return and treatment effects: The Mincer equation and beyond. In E. A. Hanushek & F. Welch (Eds.), Handbook of the economics of education (Vol. 1, pp. 307–458). Elsevier. https://doi.org/10.1016/S1574-0692(06)01007-5 (Quotation from NBER Working Paper 11544, p. 3.)
• Lev, B., & Gu, F. (2016). The end of accounting and the path forward for investors and managers. Wiley. Quotations are from (2016a) Lev, B., & Gu, F., "The End of Accounting," The Wall Street Journal, 21 June 2016.
• Radin, A. J. (2016, December 23). The end of accounting and the path forward [Book review]. The CPA Journal. https://www.cpajournal.com/2016/12/23/end-accounting-path-forward/
• Tyler, J. H., Murnane, R. J., & Willett, J. B. (2000). Estimating the labor market signaling value of the GED. Quarterly Journal of Economics, 115(2), 431–468. https://doi.org/10.1162/003355300554818

Standards, regulation, and data sources

• International Accounting Standards Board. IAS 38 Intangible Assets, paras. 4, 8, 15, 69. Text checked in the New Zealand adoption, NZ IAS 38 (External Reporting Board, January 2023), which incorporates the IASB standard: https://standards.xrb.govt.nz/assets/dms-assets/NZ-IAS-38-Jan23.pdf
• International Accounting Standards Board. IFRS 3 Business Combinations, para. B37. Text checked in the Australian adoption, AASB 3 (compiled June 2020): https://aasb.gov.au/admin/file/content105/c9/AASB3_08-15_COMPjun20_01-22.pdf
• US Securities and Exchange Commission. (2020, August 26). Modernization of Regulation S-K Items 101, 103, and 105 (Release No. 33-10825). https://www.sec.gov/files/rules/final/2020/33-10825.pdf
• International Organization for Standardization. (2025). ISO 30414:2025 Human resource management — Requirements and recommendations for human capital reporting and disclosure. ISO.
• World Bank. (2020). Human Capital Index 2020: Bangladesh (country brief). https://databankfiles.worldbank.org/public/ddpext_download/hci/HCI_2pager_BGD.pdf
• "More than 2.6m unemployed in the country: BBS." Prothom Alo (English), 29 March 2023. https://en.prothomalo.com/bangladesh/r0k3kng24p
• "Unemployed graduates double to 8 lakh in 5yrs." The Daily Star, 26 October 2023. https://www.thedailystar.net/business/economy/news/unemployed-graduates-double-8-lakh-5yrs-3453956

Kushiyara Bank, the calculator inputs and results, the comparison of the two theories, the readings of the Bangladesh data, and the implications for HR managers are original teaching material for this series, not findings reported by the sources above.

Next in The Human Capital MBA series: Article 13 continues Module 03, Human Resource Management Foundations.