The OLS and ordered probit estimation of a causal relationship between income and happiness assume linearity in individual and average income effects and the same average effect holds over the entire range of subjective well-being distribution. The response of well-being to income changes is not the same for poor and rich, dissatisfied or unhappy, and satisfied or happy people. This paper estimates heterogeneity in the income-life satisfaction relationship at specific locations of subjective well-being distribution by panel quantile regression method. The data is derived from six waves of the World Value Survey for 12 major states of India for 24 years over the period 1990-2014. The descriptive analysis of well-being distribution across states shows heterogeneity in the income-well-being relationship between states and individuals within states. The across-states over-time panel quantile regression results reveal that estimation of average effect only provides an incomplete picture of the effect of income at both ends of the conditional distribution of well-being indicators, life satisfaction and happiness levels. The quantile regression estimates vastly differ not only from the mean estimates, but also across quantiles. Life satisfaction falls and happiness rises with an increase in average state income and the income effect is strong at the lower end than at the upper end of subjective well-being distribution.