Human Capital, Education, and Economic Growth

Human Capital, Education, and MEG

• Human capital influences the quality of labor input through aspects like age, experience, education, training, and skill.
• Globalization spreads information, reduces information asymmetry, assesses profit opportunities and improves the integration of labor and capital into new economic environments which leads to adaptability and further reduces transaction costs, increases labor and capital mobility, and expands trade flows.
• Innovation and technological diffusion related to the Second Industrial Revolution require a technically adept elite and widespread literacy and numeracy among industrial laborers.
• Technological elite identifies the technological frontier and handles development, installation, and adaptation.
• Manufacturing labor requires technical skills and numeracy for operating new industrial technology.
• Adaptability is needed due to the rapid pace of technological change.
• Laborers must be literate to read instruction manuals.

Measuring Human Capital

• Basic numeracy could be measured through age heaping in census ages, tax records, death records, and legal documents.
• Basic literacy is measured through literacy outputs and inputs.
• Measuring by literacy output assumes people learn to read, then write, then count.
• Individuals signing their names suggests they are at least semi-literate.
• Evidence of signing ability can be found in marriage registers, law suits, license applications (e.g., bakers), and contracts (e.g., loans, apprenticeships), but note that signing name ≠ writing.
• More formal testing for literacy includes tests given to prisoners, migrants, and army recruits.
• Representativeness of these samples may be limited.
• Formal literacy tests are relatively rare and not random, typically involving young, urban males with higher education.
• Prisoners often have lower education but receive formal education in prison, with repeat offenders having more opportunities.
• Army recruits are from rural and urban areas but are exclusively male.
• Army recruits have an incentive to misrepresent literacy.
• Migrants' self-reported data is not carefully tested but represents a broader mix of the population.
• Literacy inputs include legislation specifying access to education, with publicly provided resources, entrance fees, entrance exams, and compulsory enrollment/attendance/enforcement.
• Government statistics on education and schools include data on the number of schools (and their size), students, teachers, and spending on education.
• Legislation and information on access only indicate "potential" for literacy/education, not successful education.
• Measuring industrial and technological elite quality is difficult, as it is determined by the ability to invent, innovate, and adapt.
• Success is not closely correlated to MEG (UK did not perform well by this metric, even before 1830).
• Potential for tech elite can be indicated by the number of engineers, college/university spots, professors, job market success of graduates, and subsidies for tech education.
• Measuring informal tech education is even harder which includes apprenticeships and on-the-job training.

Evidence on Literacy, Education and MEG

• A positive correlation links measures of human capital (literacy and numeracy) to rapid and persistent growth and MEG.
• Higher literacy correlates with higher growth.
• Links human capital to net benefits of formal education.
• Formal education links to government involvement through subsidies, public provision, child labor laws, or compulsory education laws.
• Government education expenditure links to formal education, literacy and numeracy, and growth.
• Countries moving up the global ranking of economic performance have high and rising literacy rates.
• The speed of convergence to the UK after 1850 is closely and positively correlated with literacy rates and rank correlation between literacy and growth.
• Causation suggests faster-growing countries in the 19th century had the highest literacy rates.
• The distribution of literacy within countries can reveal causal connections.
• Literacy is not uniformly distributed, variance across groups within countries is > variance in (young-urban-males) across countries.

Adult Illiteracy Rates

• Germany (1871):
  • Protestant men = 7%
  • Protestant women = 11%
  • Catholic men = 15%
  • Catholic women = 22%
• England (1851):
  • Urban men = 13%
  • Urban women = 23%
  • Rural men = 43%
  • Rural women = 45%
• US (1850):
  • "White" men and women = 10%
  • "Black" men and women = 80-90% (estimated because nobody bothered to count)
• Literacy rates are generally higher for males than females, urban than rural, rich than poor, industrial than agricultural, young than old, Protestants than Catholics, and "white" communities than racialized communities.
• Differences in literacy and education across countries and groups may reveal differences in MB and MC of education, following Becker's rational decision-making model.
• O'Rourke and Williamson (1999) tested the empirical relationship between literacy (education) and the late 19th-century convergence rate with statistically significant results for "Beta" and "gamma".
• A 1% literacy rate greater than average in 1870 would have seen approximately 0.5% faster convergence.
• "Deviant" literacy country-specific percentage include: Sweden = 31%, US = 21%, Germany = 0%, Portugal = 75%, Spain = 60% and Italy = 100%.
• Acquiring basic literacy increases the rate of convergence while more specialized education beyond basic literacy has little to no impact.
• Large differences in literacy drive differences in convergence, but smaller differences don't matter nearly as much.
• Differences within a "literate" group don't matter, and within an "illiterate" group differences don't matter, but differences between literate and illiterate groups are explained by differences in literacy.
• Literacy cannot explain the timing of US-UK convergence, but German and Scandinavian literacy increases coincide with sharp growth.
• Changes in literacy are not closely correlated to transitions to MEG in many converging economies.
• Government policies can influence private families' decisions regarding children's education, hard to measure the private families' marginal benefit (MB).
• Government education policies are correlated with formal education, literacy, numeracy, and growth.
• Governments spending more on subsidizing primary education had more children in school, higher literacy rates, and faster growth rates during the late 19th century.
• This is linked to government subsidization, primary enrollment, literacy, growth and MEG.
• The causal mechanisms at work, and the role of government policy are factors to consider.
• Governments that don't provide education is also a factor to consider.

Education as a Rational Economic Decision

• Becker's economic decision-making model applied to education asks, "Should children stay in school this year, or drop out?"
• Individuals should stay in school if the expected present value of the net benefit of an additional year is positive.
• The net benefit is calculated using the formula Maxy EPV(TB) – EPV(TC) with the constraint s.t. Incomet ≥ TCt ∀ t.
• EPV TB (Expected Present Value of Total Benefits) is the expected present value stream of increased income due to one additional year in school.
• The EPV stream includes other welfare benefits from education.
• Income change in direct relation to a year in school is not guaranteed to be continuous, linear, or even positive.
• Reading (earliest years schooling) increases future income considerably.
• Writing increases future income slightly less.
• Counting (later years) increases future income even less.
• Specialized skills may increase income at the top of the income distribution, but their impact on EPV TB may be limited (MB<0).
• Specific outcomes increase the potential income but the incremental improvement has little impact.
• The education without specific outcomes has a small effect on income measures.
• Uncertainty about payoffs increases when payoffs are in the future and technological change is rapid which implies ↑ risk leads to ↓ EPV TB.
• Time discounting also ↓ EPV TB for income earned in the distant future.
• The TB ↑ during MEG which is the marginal benefit (and ↑ as rate and complexity of tech A↑).
• EPV TC (Expected Present Value of Total Costs) is equal to the total current and opportunity costs to stay in school.
• Tuition, books, accommodation, food, and clothing is among the components.
• Foregone income from students not working.
• Relative to TB, costs are current and certain (no risk or time discount).
• MEG generates urban/industrial employment opportunities for young males, which increases the opportunity cost and foregone income from education.
• Choosing the optimal number of years of education can be described as a constrained maximization problem with MB = MC at Y*.
• The model illustrates endogeneity in the decision process: education can cause MEG and vice versa.
• MEG raises the returns (MB) of education (↑ income + ↑ probability of employment + ↓ uncertainty + ↓ time lag to payoff).
• BUT unskilled labor is complementary to industrial capital and skilled technological elite labor, which is why return increases to relatively unskilled labor.
• Skill premium shifts upward during MEG.
• Nations with favorable social/cultural institutions and policies have a higher NB of education, which leads to a higher level of education, or Y*.
• Government subsidies decrease private individuals' MC, while compulsory schooling and laws decrease school labor.
• Social mobility and cultural attitudes that value education will increase the marginal benefit.

Political Economy of Education

• If formal education drives transitions to MEG, and governments can incentivize families, then why don't more governments subsidize formal education?
• Lindert (2004) provides a model to explain patterns in subsidization for education.
• Education has positive externalities: it benefits private individuals who pay for the costs to fund education.
• Private individuals receive MB from education.
• Society also gets some benefit from private education decisions like globalization, technological advancements, and MEG.
• Private individuals choose Y* to maximize their private NB, not society's NB which leads to too little education being chosen because private individuals ignore positive externality of their decision.
• Private individuals optimally choose Y* years of education, but social welfare max level of education takes into account the increased return on investment in externalties (ie. SMB-PMB) which will push the graph to Y**.
• Benevolent social planners should aim to lower PMC so private individuals optimally choose Y**.
• Governments are not always benevolent social planners and may respond more to political economy incentives than to social welfare.
• Lindert explains how education subsidization happened in UK, US and Germany in the late 19th century .
• Education is expensive, requiring high income/wealth/land taxes to pay for formal education.
• Special interest groups can influence government subsidization of education.
• Land owners are against education taxation, viewing unskilled labor as complement and skilled labor as a substitute which lowers return on land.
• Religious elites are split on education, depending on whether they are Protestant or Catholic.
• Bureaucracy can favor taxes on education as it requires sophisticated, complex processes, such as taxation.
• Skilled labor can go either way as it garners highest incomes and creates competition to current skilled workers.

Conclusions

• Government subsidization is more likely in countries with poor, disorganized, politically inactive landowners, wealthy, concentrated K owners, competing disorganized religious elites, a large, sophisticated, educated bureaucracy, and disorganized skilled L.
• The theory almost perfectly describes the political economy in the US in the early 19th century, which is the opposite of BOTH the UK and Germany.
• The US was the first to subsidize both elementary and high school education fully.
• The UK only started subsidizing elementary schools in the 1890s.
• Germany was a patch-work.
• At the national level, German special interest groups were more closely aligned with UK groups.
• The German regions split between Protestant and Catholic, giving birth to strong bureaucracy and organized unions.
• These episodes also contributed to the decentralization of German education policy, and this is largely set at local, not central, levels.
• There will be a distribution of voter preferences where "median voter theorem" is concerned on any issue related to schools.
• Decentralization will shift the decision closer to the median voters.
• Patch-work reflects regional split in education.
• To achieve education equality, countries needs the special interest groups to align with the media voters.

The US Model of Human Capital Provision

• The US had the highest literacy rates very early in the 19th century, and the US model had a good MB.
• The US education system was described by publicly funded elementary and high school education, general human skills transferable across markets, and decentralization of the funding system.
• British Education model was in sharp contrast, only subsidizing elementary schools by the 1890s, with classical education, and only some of the basic skills taught in the US.
• Goldin describes the demand in schools and how that drove the specific features of the US and wealth taxes.
• Special interest groups favoring education include capital owners who are wealthy and politically open.
• In the US, US technological and rapid change in the system made mass education important, as well as mass participation through the population when voting.
• High return to human capital is very high in this system, complex because opportunity cost is low to the US unskilled labor.
• Wealth accumulation and practical/applied education has worked to incentivize human mobility.

Conclusion

• Decreased skill premium supply by the demand in schools and wage gaps is something they aim to get rid of.
• Wage inequality can be curbed by having education go through a greater spectrum.
• US models have been very successful in integrating various skills such as low premium with L mobility.
• Hard to prove what the causation has been, a long question to solve.
• In Iowa by 1915 the lower wages are being combated by incentivizing L.
• In Iowa and Ohio, wage workers get the same workers/clerical pay to rise.

Human Capital and Education Conclusions

• Higher subsidized education is correlated with growth during the 19th century.
• Theoratically MEG COULD CAUSE education or it can go either way.
• Basic literacy is very important, and countries with the basic fundamentals create higher levels of education.
• US has been doing it well.