The present study is an attempt to examine the trends in returns to education in light of the long-term economic growth in India during 1983 to 2009-10. It outlines various forms of inequality issues prevalent in Indian labour markets, with respect to the rural/urban areas, gender, caste and nature of work. The unit level data from 6 rounds of National Sample Survey during 1983, 1987-88, 1993-94, 1999-2000, 2004-05 and 2009-10 were used for this study. Mincer wage function was estimated by using the OLS method and the results were also compared to the median wage equation, which proved the consistency of these estimates. The casual wage markets for males provided incentives for higher education till some intermediate levels in the form of higher wage earnings than their illiterate or below primary educated counterparts but no additional advantage for secondary or graduate levels of education. Higher education could not translate into better wage earnings for female casual workers. The returns to all education levels were converging at low levels with the returns for secondary and graduate levels for urban casual male workers declining over time. There was a decline in the returns to secondary and graduate level of education for rural male regular workers with almost no change in the pattern of returns for urban male regular workers. The returns to education for graduation for female workers increased tremendously due to increased employment opportunities for better educated females in the India during the last decade of fast economic growth, led largely by the growth of the service sector. While there is need to enhance public investment in education for improving higher education opportunities in India, there is also a need to reorient rural education by focusing on imparting working skills between middle level of education and secondary levels. The education curriculum must ensure that higher education translates into better wage earnings for the unskilled or semi-skilled majority of the rural workforce in the long run.