In this paper, we take a closer examination to the so-called natural hedging approach proposed for life insurers to internally hedge their systematic longevity risk exposures. Relying on a non-parametric mortality forecasting model, we manage to detour from the debated factor model assumption and conduct a more realistic approach to test the actual effectiveness of natural hedging. In particular, our numerical results of calculated optimal economic capitals suggest that for life insurers, natural hedging may not be as effective as proclaimed by the existing insurance literature in handling longevity risk.