There is considerable controversy regarding the long-term strength of continents (T _e ). While some authors obtain both low and high T _e estimates from the Bouguer coherence and suggest that both crust and mantle contribute to lithospheric strength, others obtain estimates of only <25 km using the free-air admittance and suggest that the mantle is weak. At the root of this controversy is how accurately T _e can be recovered from coherence and admittance. We investigate this question by using synthetic topography and gravity anomaly data for which T _e is known. We show that the discrepancies stem from comparison of theoretical curves to multitaper power spectral estimates of free-air admittance. We reformulate the admittance method and show that it can recover synthetic T _e estimates similar to those recovered using coherence. In light of these results, we estimate T _e in Fennoscandia and obtain similar results using both techniques. T _e is 20–40 km in the Caledonides, 40–60 km in the Swedish Svecofennides, 40–60 km in the Kola peninsula, and 70–100 km in southern Karelia and Svecofennian central Finland. Independent rheological modeling, using a xenolith-controlled geotherm, predicts similar high T _e in central Finland. Because T _e exceeds crustal thickness in this area, the mantle must contribute significantly to the total strength. T _e in Fennoscandia increases with tectonic age, seismic lithosphere thickness, and decreasing heat flow, and low T _e correlates with frequent seismicity. However, in Proterozoic and Archean lithosphere the relationship of T _e to age is ambiguous, suggesting that compositional variations may influence the strength of continents.