Control design of inverter-based microgrids plays a significant role in affecting dynamic performance of the system. Conventional microgrid droop control suffers from instability due to low X/R ratios and unique network characteristics as compared to large power systems. While many approaches such as virtual framework methods, virtual impedance methods, or synchronverters have been proposed and proven effective, an intuitive and fundamental insight into physical origins of instability has not yet been sufficiently disclosed. In this paper, a systematic approach for enhancing the stability of inverter-based microgrids is proposed. A test system is studied to derive simple and concise stability criteria based on the proposed Lyapunov function method. Particularly, we show that unlike in large-scale power systems, for microgrids the transient susceptance B′ plays a crucial role in contracting the region of stable droop gains. Control schemes to minimize B′ are then investigated, enabling a different perspective in views of the virtual component method. Finally, simulations are carried out to validate the proposed approach via direct time-domain analysis.