In this section, we develop a theoretical model to study the link between globalization and public expenditures for different educational stages. More specifically, after setting up the basic structure of the model in sections 4.3.1–4.3.4, we explore in section 4.3.5 how global-ization affects the endogenous variables in the model, i. e. public education expenditures and the domestic tax rate.
4.3.1 Individuals
Consider a country with a population mass of 1. An individual i has an exogenously given ability that theoretically qualifies her for one and only one type of labor, for example “high-skilled” or “low-“high-skilled” work. The wage that this individual earns for one unit of effective labor iswi. The effective labor supply of individualidepends on the amount of public expen-ditures gi that the government invests in her education. Public education expenditures are hence assumed to be productivity-enhancing. Individuali’s market incomeνi is consequently specified asνi(wi, gi) withdνi/dwi >0,dνi/dgi >0, d2νi/dg2i <0, andd2νi/dgidwi >0.
These assumptions imply (i) that increasing wages and an increasing effective produc-tivity due to more funding for the relevant educational stages raise the market income of individuali; (ii) that education expenditures have a declining marginal effect on income; and (iii) that the marginal effect of education expenditures on income rises with higher wages.
The idea behind these assumptions is that the ability and talents of the individuals in the model are not substitutable. An individual with academic abilities can only pursue an academic career, whereas an individual with practical skills can only work in “practical jobs”.
However, possessing the respective abilities is not sufficient. Individuals have to receive an appropriate education before their talent can be productively applied. If they receive either no education at all or the wrong kind of education, they become unproductive. For example, an individual with innate academic talents will become completely unproductive if she does
not receive higher education (it cannot take a practical job instead). Similarly, an individual with practical skills will benefit from good primary and/or secondary secondary education, but does not benefit from higher education.
Emigration
One important constraint the government faces when formulating its fiscal policy is that individuals may emigrate if the tax burden is too high.7 To model the mobility decisions, we presume that every individual takes the tax rate into account when deciding whether to emigrate or not. Individuals will remain in the home country if the following condition holds:
(1−t) +i ≥(1−tF)−x, (4.1)
with tF denoting the tax rate in case of emigration (the “foreign” tax rate),x denoting the costs that have to be incurred in the case of emigration, and i a random parameter that measures the home attachment of a given individual. We assume that i ∼ U(0,1), i. e., that home attachment is uniformly distributed over [0,1]. An individual will emigrate if the difference between domestic and foreign tax rates is larger than her home attachment and the mobility costs.
Given thatiis random, every individual’s mobility decision is stochastic. The probability πithat an individual will remain in the country can be expressed as a function of the domestic tax rate and the mobility costs:
πi=πi(t, x) =F(i≥z) = 1−z, (4.2) with z= (t−tF)−x,dπi/dt=−1, anddπi/dx= 1.
4.3.2 The government
We follow Anderson and Konrad (2003) by modeling the government as a Leviathan. The government is therefore exclusively concerned with rents R, which are defined as tax receipts minus total expenditures for education. The objective function is:
maxgi,t R= Z 1
0
(tπiνi−gi)di (4.3)
with gi≥0,0≤t≤1.
7Assuming that the production factors are supplied endogenously would lead to an alternative tax base effect. We ignore this effect in order to keep the model tractable.
Thus, education expenditures that increase individuals’ incomes are only of interest to the government as far as they lead to higher rents.
4.3.3 Equilibrium
Education expenditures and the tax rate are determined simultaneously in the model. The former are characterized in equilibrium by:
tπi
dνi
dgi −1 = 0 ∀i. (4.4)
Thus, the government chooses education expenditures for every individual i such that the increase in expected tax revenues due to a marginal increase in education expenditures is equal to the costs, which are 1.
The equilibrium tax rate is determined by:
Z 1 0
πiνi+tdπi
dt νi
di= 0. (4.5)
This equation states that the tax rate is chosen such that in equilibrium the additional rev-enues due to the marginal increase in the tax rate are equal to the revenue loss due to emigration.
4.3.4 Economic effects of globalization
We now analyze how globalization affects the equilibrium characterized by equations (4.4)–
(4.5). Prima facie, globalization is assumed to have two direct effects in this model. On the one hand, it affects wages for different skill-types. On the other hand, it reduces the costs of mobility. As stated previously, we refer to these effects as “economic” effects. By affecting wages and the costs of mobility globalization will also indirectly affect the tax rate and education expenditures that the government chooses in equilibrium, i. e. it will eventually have fiscal effects.
Globalization and wages
Trade theory suggests a link between the extent of economic integration and factor returns.
The Heckscher-Ohlin model and the related Stolper-Samuelson theorem, for example, state that falling trade restrictions lead to an equalization of factor prices through an increase in the trade of goods (Krugman and Obstfeld, 2005).
We therefore model the wage of individual ias a function of globalization G:
wi =wi(G). (4.6)
How does globalization affect the wage for individuali, i.e. what is the sign ofdwi/dG? Ac-cording to the Heckscher-Ohlin model, the sign of this expression depends on (i) the skill level of individualiand (ii) whether she lives in a developing or industrialized country. Given that industrialized countries are relatively abundant in high-skilled labor and developing countries have a relative abundance in unskilled labor, one prediction of the Heckscher-Ohlin model is that the returns to low-skilled labor increase in developing and decrease in industrialized countries with deepening globalization, and vice versa for high-skilled labor. The empiri-cal evidence, however, confirms the predictions of the Heckscher-Ohlin model only partially.
That is, globalization has led to a relative rise in wages for high-skilled labor in industrialized (Feenstra and Hanson, 1999)and developing countries (Goldberg and Pavenik, 2007).
Globalization and mobility costs
The mobility costsxcan be understood as the monetary representation of the costs of losing contact or keeping in touch with one’s social and professional networks, and as the costs of relocating physical assets. One effect of globalization is that it lowers transportation costs, which implies that it becomes easier to visit one’s acquaintances in the home country, or to relocate physical assets. Another effect is the spread of English as a modern Lingua Franca and the emergence of a global culture, both of which might reduce the non-monetary costs when moving to a foreign country. It is therefore reasonable to assume that mobility costs are a decreasing function of the extent of globalization, i. e.,x=x(G) withdx/dG <0.
4.3.5 Fiscal effects of globalization
By implicitly differentiating the system of equations given in (4.4)–(4.5) with respect to G, we can analyze the effect of globalization on education expenditures and taxation. After rearranging, we obtain8:
Equation (4.7) describes the effect of globalization on education expenditures and equa-tion (4.8) its effect on the tax rate. As argued previously, equaequa-tion (4.7) reveals that globaliza-tion affects educaglobaliza-tion expenditures for individualithrough two channels: (i) by affecting the wage for the type of labor that individualisupplies (dwi/dG) and (ii) by affecting mobility costs (dx/dG).
The wage effect varies between individuals. In particular, the sign of dwi/dGwill likely differ between high-skilled and low-skilled individuals. As argued previously, the available em-pirical evidence indicates that dwi/dGis positive for high-skilled and negative for low-skilled individuals in both industrialized and developing countries. Thus, this effect of globalization provides governments with an incentive to increase expenditures for higher education and reduce expenditures for lower education in both developed and developing countries.
The effects of the declining mobility costs are the same for all i. First, it is easy to see from equation (4.7) that the reduction in mobility costs motivates the government to reduce expenditures for all educational stages since dx/dGis negative.
However, the overall impact of globalization on education expenditures depends on how it affects the tax rate as well. But the tax rate is an endogenous variable so that the sign ofdt/dG has to be determined within the system described by equations (4.7)–(4.8). Unfortunately, it is neither possible to explicitly solve equation (4.8) for dt/dG, nor to determine whether this expression will be positive or negative.
More specifically, globalization has three effects on the domestic tax rate. First, it in-centivizes the government to lower the tax rate to prevent emigration because of the smaller emigration costs (dx/dG <0). Second, it affects national income through two channels. One the one hand, it has an effect on the wages of all individuals. On the other hand, it affects incomes indirectly through its effect on education expenditures (which is determined within the system). It is unclear how the effect on income will impact the tax rate in the new equilibrium. At a fundamental level, we cannot be even sure that globalization’s effect on national income will be positive or negative in a given country.
But even if we are willing to assume that globalization increases national income in a given country, the government faces conflictive incentives regarding taxation. On the one hand, it has an incentive to increase taxation because the marginal returns for a one percentage point increase in taxes are higher if income becomes larger. On the other hand, the costs in terms of forgone rents are also higher if individuals decide to emigrate because of tax rate differentials.
Thus, whether or not governments will increase or decrease the tax rate due to globalization cannot be determined in general. Rather, the government’s decision will depend on the values of the model parameters at a particular equilibrium.
While this model hence establishes that there is a link between globalization, the equi-librium tax rate, and different types of education expenditures, it offers no unambiguous prediction regarding the sign of the effects. That is, because of its ambiguous effect on the
tax rate, the impact of globalization on absolute education expenditures for individual iis, irrespective of her skill-type, ambiguous as well. The discussion on the effect of globalization on the wage for different skill-types of labor, however, tentatively suggests that spending for higher eduction should increase relative to spending for lower education with deepening glob-alization in both developed and developing countries. This, then, is a theoretical hypothesis to be tested empirically.