The rapid shift to inertia-less variable renewable energy (VRE) technologies is displacing many synchronous generators (SGs), reducing frequency support and inertial response, and significantly affecting system frequency dynamics. Therefore, at high levels of VRE penetration, it is crucial to accurately represent frequency stability constraints in power system operation and planning. This paper presents a two-stage methodology to assess the impact of frequency stability constraints in the generation expansion planning (GEP) model, focusing on VRE integration and the dynamic performance of power systems. The first stage involves the GEP model determining the optimal new generation capacity mix, considering linearized frequency stability constraints. The second stage assesses the GEP results' dynamic performance by simulating power system transients using DIgSILENT PowerFactory. To quantify the impact of frequency stability constraints, we consider three security indices: initial rate of change of frequency (RoCoF), frequency nadir, and steady-state frequency deviation. Finally, we analyze two case studies: the GEP model with and without considering frequency constraints on an illustrative 13-bus test system. The results demonstrate that integrating frequency constraints into the GEP problem results in higher system inertia, leading to significant improvements in the security indices but a more expensive investment in generation portfolios.