For the purpose of explaining both business cycles and asset returns, we examine a real business cycle (RBC) model with habit-augmented preferences and endogenous costs of adjusting the capital stock. Following the agency-cost model of Carlstrom and Fuerst (1997), capital adjustment costs are affected by the level of entrepreneurs net worth such that an increase in net worth (following a positive productivity shock) lowers agency costs associated with external financing, and hence makes it easier to expand the capital stock. Along with the restricted labor supply, the model resolves the asset pricing puzzles of the consumption-based model in the sense that the implied stochastic discount factor (or pricing kernel) reaches the Hansen-Jagannathan (1991) volatility bound. Further, this improvement in the asset pricing dimension is achieved without reducing its business cycle performance such as output and consumption volatility. This is in a sharp contrast to the standard RBC model with the reduced-form adjustment cost technology where sufficiently low supply elasticity of capital (or persistently high capital adjustment costs) is required to generate the equity premium at the expense of low output volatility. Here, the capital supply is highly elastic with respect to Tobin's q under the plausible calibrations of the structural parameters affecting endogenous capital adjustment costs. The sluggish behavior of net worth, as a shifter of the capital supply curve, is the key mechanism by which capital adjustment is delayed, hampering consumption smoothing desired by households with habit persistence preferences. The agency-costs model reveals that a small curvature in the capital adjustment cost function, viewed as crucial for understanding the fluctuations in Tobin's q, can be also consistent with both the historical equity premium and the key business cycle facts.