This chapter presents the application of econometric methods for productivity analysis. Leveraging the duality between cost and production functions, the chapter addresses endogeneity and lag issues, facilitating estimation of TFP and factor elasticities. It introduces traditional, restricted, and generalized cost functions, illustrating parameter estimation, hypothesis testing, and TFP decomposition. For traditional cost functions, linear homogeneity and consistency between cost and cost-share equations are imposed, with 3SLS used for parameter estimation. The restricted cost function handles unavailable input prices using instrumental variables to address endogeneity. Shadow prices are derived via the Hotelling Lemma. The generalized cost function, due to its nonlinearity, is estimated using the nlsur command through feasible generalized nonlinear least squares. Practical Stata code examples demonstrate variable construction, constraint implementation, and calculation of TFP growth, its driving factors, and shadow prices, providing a stepwise guide for applied econometric programming in productivity and cost analysis. Cost functions and production functions share a dual relationship, a connection that effectively addresses issues such as endogeneity and lag inherent in production functions. This relationship is of great significance as these functions are indispensable in estimating crucial metrics, including total factor productivity (TFP) and factor elasticity. In this section, we'll provide a concise introduction to several key aspects. Firstly, we'll look at how to use software applications when solving productivity problems with the cost function and restricted cost function. Secondly, we'll cover the estimation of substitution elasticity and shadow prices. Lastly, we'll explore the generalized cost function used for productivity estimation.