The elasticity of substitution (ES) between capital and labor is a crucial parameter in the literature of economic growth, especially in analyzing functional income distribution. When we look at the literature, we will see that estimations of the ES depend on the structure of data, assumptions on the functional form, and choice of the estimation method. This study departs from the Cobb-Douglas assumption of unitary ES between capital and labor and estimates the ES from input demands by using the constant elasticity of substitution (CES) production function. The data used in the study is the Annual Industry and Service Statistics Micro Data Set which is obtained from TURKSTAT and it covers the period from 2003 to 2015. This data set enables us to obtain different ES parameters for each industry. Therefore, an industry-level analysis helps us to avoid biases resulting from aggregation and allows us to interpret the substitutability of production factors by considering the characteristics of industries. After employing fixed error panel data method with time trend for technological change, preliminary regression results revealed that the ES is below unity in all industries.