This thesis consists of three empirical studies in economics of education on the determinants and consequences of language-in-education (LiE) policies. The “Environmental settings – Inputs – Processes – Immediate outcomes – Long-term outcomes” (EIPOL) evaluation model is applied to LiE policies and programs and serves as the overall framework of this research (see Introductory Chapter). Each study then targets at least one stage of the EIPOL framework to test the validity of the “green” vs. “free-market” linguistic theories. Whereas the two first studies derive models tested empirically in the African context, the third is tested on a sample of countries from the International Adult Literacy Survey (IALS). The first study, Rationales to Language-in-Education Policies in Postcolonial Africa: Towards a Holistic Approach, considers two issues. First, it explores the factors affecting the choice of an LiE policy in 35 African countries. The results show that the countries adopting a unilingual education system put different weights on the influential parameters than countries adopting a bilingual education system and that both groups of countries validate somehow both the “green” and the “free-market” approaches. Second, the article investigates how decision makers can ensure the optimal choice of language(s) of instruction by developing a non-cooperative game theoretic model with network externalities. The model shows that it is never optimal for two countries to become bilingual, or for the majority linguistic group to learn the language of the minority group, unless there is minimum cooperation to ensure an equitable redistribution of payoffs. This finding confirms the “free-market” theory. The second study, The Role of Language in Learning Achievement: A amibian Case Study, investigates the role played by home language and language proficiency on mathematics scores of 5048 Grade-6 learners in 275 Namibian schools, via the second survey data by the Southern and Eastern Africa Consortium for Monitoring Educational Quality (SACMEQ). Hierarchical linear modeling is used to partition the total variance in mathematics achievement into its within- and between-school components. Results of the analysis show that although home language plays a limited role in explaining within- and between-school variations in mathematics achievement, language proficiency, when proxied by reading scores, plays a significant role in the heterogeneity of results. Thus, confirming the role of language skills in learning achievement and so validating the “green” theory. Finally, the third study, Language Skills and Economic Returns, investigates the economic returns to language skills, assuming that language competencies constitute key components of human capital. It presents results from eight countries enrolled in the International Adult Literacy Survey (IALS). The study finds commonalities between countries in terms of the valuing of language skills, independent of the type of language policy applied at the national level. In each of the eight countries compared, skills in a second language are estimated to be a major factor constraining wage opportunities. This study validates the “free-market” theory.