Parental Quality Investment, Child Effort, and Human Capital Accumulation

Abstract

This paper develops a theoretical model that explores the effects of child investment on human capital accumulation. Household investment effects are measured as a function of quality of investment choices, time allocation, child effort, and spillover effects. The theory dynamics are derived by employing a two-period OLG model. The model dynamics reveal the existence of multiple steady states of quality investment and child human capital accumulation under two stages of development. These dynamics show why poor households are often unable to escape poverty. Empirically, the theory and dynamics are tested with United States data. The data suggest that for poor households, income and parental human capital have no significant effect on child performance at school. Child effort and parental quality investment are shown to be significant determinants of child performance at school across household types.

Keywords:

• Child investment
• child effort
• human capital

JEL CLASSIFICATIONS :

• E13
• H23
• H31
• O11

Acknowledgements

This paper benefited from interactions with Arthur Denzau, Thomas Willett, Simon Lamar, Paul Zak, Amit Ghosh, anonymous referees, and countless others. This work is an extension of my dissertation and as such I am indebted to Claremont Graduate University. As always, all remaining errors are my own.

Notes

¹High levels of human capital countries experiencing stagnating growth.

²Path dependent outcomes.

³This result is significant for children of mothers who are enrolled in school. Currie and Morretti (2003) provide a nice empirical analysis of the effects of mother's education on child development.

⁴Where 1 is poorer than 2 and 3 is richer than 2.

⁵This is consistent with the work by Cipriani and Makris (2007), who allow child effort to evolve positively at a decreasing rate. Note that Cipriani and Makris (2007) define child effort as an increasing function of time spent at school.

⁶Note that Levy (2003) uses the term ‘talent’, rather, than ‘ability’. In this model, it is assumed that talent and ability are consistent with endowed birth factors.

⁷The author thanks an anonymous referee for this suggestion.

⁸This commitment, is referred to as parental nurture.

⁹The first functional form assumes that ability is not determined by parent human capital. Work by Maoz and Moav (1999) similarly assumes independence between ability and parent human capital. The second assumption is similar to the setup in Becker and Tomes (1979) that assumes a constant fraction transmission to child. Work by Becker and Tomes (1986) also uses a similar setup in determining child endowment.

¹⁰It stands to reason that for values of ω<1, the investment term, θ, will have a decreasing returns effect on h _{t+1, j} . The use of ω on household investment choice is consistent with the set-up used by Tamura (2006). Note that Tamura (2006), using his notation, assumes that ρ>0 in its effect of parental rearing time. The assumption that ω>0 is consistent with his computational exercise that found his ρ=1.02, 0.92.

¹¹Alternatively, ω can be linked to output productivity, in which case, along the same lines as Doepke (2004), for sufficient high productivity in industry there will be a switch from a Malthusian regime to a growth regime.

¹²There are two underlying assumptions being presented here. First, kids with higher grades are on average kids with higher human capital relative to those who have lower grades. Secondly, kids with higher grades are, on average, more likely to attain, or attempt, higher levels of education.

¹³Llozano (2008) looks at school leadership roles on college completion. Jepsen (2008) finds that 8th grade test score is a positive and significant predictor of college attendance. Similarly, Ganderton and Santos (1995) find both test scores and high school GPA as a strong predictor of college attendance and college completion.

¹⁴ h _t is taken from the responses of mother and father highest level of education.

¹⁵The subjective breakdown is one unit above that chosen by the authors of the survey. Please, see Table 2 for a description of these and all other variables used. Note that all small cap variables are modifications of the original variables that appeared as upper case variables in the dataset.

¹⁶ mommths and dadmths ask for the number of months she or he spent working for pay or income in the past 12 months, see Table 2.

¹⁷It is possible that a parent's time can be allocated to a third activity that is not measured by this proxy. Say, work, rearing child, and drinking with buddies. At best, it can be assume that time spent rearing child is a constant fraction of all non-working time. At worst this measure is a control variable of known time spent away from home. Nonetheless, the omission of these two variables does not alter the results that follow.

¹⁸The outcomes (1), (2),˙s, (4) represent the likelihood of being a ‘D’, ‘C’, ˙s, ‘A’ student, respectively.

¹⁹However, this result is sensitive to the income category used. When the measure of each groups household income is adjusted to that used by the authors of the survey it makes mother's time at work have a significant impact on middle income households. In addition, it takes away the significance found on high income households. See Table 7.