[en] We measure the merger fraction of Type 2 radio-loud and radio-quiet active galactic nuclei (AGNs) at $z><!--\; >\; -->1$ using new samples. The objects have Hubble Space Telescope (HST) images taken with Wide Field Camera 3 (WFC3) in the IR channel. These samples are compared to the 3CR sample of radio galaxies at $z><!--\; >\; -->1$ and to a sample of non-active galaxies. We also consider lower redshift radio galaxies with HST observations and previous generation instruments (NICMOS and WFPC2). The full sample spans an unprecedented range in both redshift and AGN luminosity. We perform statistical tests to determine whether the different samples are differently associated with mergers. We find that all (92%${}_{-<!--\; ???\; -->14\mathrm{\%<!--\; \%\; -->}}^{+8\mathrm{\%<!--\; \%\; -->}}$) radio-loud galaxies at $z><!--\; >\; -->1$ are associated with recent or ongoing merger events. Among the radio-loud population there is no evidence for any dependence of the merger fraction on either redshift or AGN power. For the matched radio-quiet samples, only 38%${}_{-<!--\; ???\; -->15}^{+16}$ are merging systems. The merger fraction for the sample of non-active galaxies at $z><!--\; >\; -->1$ is indistinguishable from radio-quiet objects. This is strong evidence that mergers are the triggering mechanism for the radio-loud AGN phenomenon and the launching of relativistic jets from supermassive black holes (SMBHs). We speculate that major black hole (BH)–BH mergers play a major role in spinning up the central SMBHs in these objects.