A contrast between the outcomes in Hungary and Romania is puzzling. Despite the barbarity of the Romanian authorities, approximately half of Romania’s Jews survived, a larger percentage than in Hungary. Out of 756,000 Jews in Romania in 1930, 375,000 survived the war, the vast majority of them in Regat Romania [the old kingdom of Moldavia and Wallachia, not the parts of Greater Romania acquired after World War I]. Why? There are essentially two reasons for this outcome, both consistent with the theoretical framework put forward here emphasizing losses at the outset. First, as Radu Ioanid put it,
In regard to the experience of the Jewish community of Regat, one thing was clear during the Holocaust: not having come into contact with the Soviets in 1940, the Jews were not held accountable for the loss of Bessarabia and Northern Bukovina and therefore not singled out for prompt punishment at the beginning of the war.
Thus, Jews in the Regat were not murdered in the same extent as those in Bessarabia and northern Bukovina, because they were not held responsible for the Romanian territorial losses and consequent refugee migrations.
Second, despite Romanian refusals to initiate these deportations, the Germans might still have intervened directly in Romania to effect deportations under different geopolitical circumstances. And here we find one of the crucial distinctions between Romanian and Hungarian behavior. Whereas the Romanians could refuse the German requests for deportation for their own reasons, having to do principally with the fear of Allied reprisals after the increasingly likely German defeat, the Hungarians could not. In contrast to Romania, Hungary lay directly in the path of the Soviet westbound march. In the Nazi view, as we saw earlier, the large concentrations of Jews in Hungary constituted a potentially collaborating fifth column that could ease the Soviet advance to the Reich heartland. Hence, direct German intervention was required.
Although geopolitically important principally due to the Ploesti oil fields, Romania did not lie directly in the path of the main Soviet advance and was not required for a strategic defense of the eastern reaches of the Reich. At this stage in the war, after Allied bombing of the oil fields and the absence of a perceived direct strategic threat Germany of Romania’s remaining Jews, an intervention was not required for strategic defense. It is ironic that a country with a far more virulent and barbaric anti-Semitic tradition could save a larger percentage of its Jews than one with an earlier history of strong Hungarian-Jewish collaboration. Yet here we see the importance of geopolitical imperatives, an important component of realpolitik (as identified in the three models [brute-force imprudent, prudent, and cynical] of realpolitik in chapter 5)….
Finally, the pattern of Hungarian-Jewish deportations suggests a transition even within imprudent brute-force realpolitik. Whereas the choice of genocidal behavior clearly was imprudent at the start of Operation Barbarossa in mid-1941, three years later, even to German opponents of Nazism, it could now appear to be prudent. By this time, the Germans could reason, many Hungarian Jews would have heard of the genocide elsewhere in Europe and would have become determined opponents of the Nazi regime. Aid to the oncoming Soviets would have been forthcoming. Having created this body of potential fifth columnists by their own unbridled brutality, the Germans were forced to live with the consequences. Deportation and death of this Jewish population then could easily have been seen by the Germans to be absolutely required in order to protect the German state and its population from Soviet revenge.
SOURCE: The Killing Trap: Genocide in the Twentieth Century, by Manus I. Midlarsky (Cambridge U. Press, 2005), pp. 258-259