The Roman registration of Herod's Kingdom is historically bizarre, and could not have been part of any "worldwide" registration. For Caesar to order a mass registration of all Roman Provinces and Client Kingdoms, at once? There is no cause to believe that. No, not even if we accept Luke 2:1-5. And personally, my desire is to trust every word of Luke 2:1-5. But what DID Augustus decree?
Regarding tax registration, Augustus' only known policy change that actually did go "worldwide" was a key facet of his provincial reforms. After 27 BC, the Princeps made sure that Proconsuls sent to govern the Provinces would be less likely to overtax their poor local subjects. Prior to that reform, the oversees Governors had normally held auctions for the job of chief tax collector, and the winning bidder had typically gone on to extort, steal or to outright seize as much "tax" as needed to make up the bid value, and then some!
Naturally, this "tax farming" system encouraged provincial unrest and potential uprisings. Wisely, Augustus decided it was finally time for Rome to start counting heads OUTSIDE OF ITALY! That is, before 27 BC, Roman Hegemony had not included the registration of foreign tax payers. After 27 BC, Rome's Provinces were now eligible for census-style registration.
However, for client kingdoms, like Herod's, the amount of the annual tribute was a matter negotiable only between Rome and the King, and this fact did NOT change after 27 BC. Clearly, Augustus was less concerned about whether some client-King's subjects were being over-taxed. To the contrary, revolts against such Kings could at times lead to Rome's occupying the kingdom, seizing the King's assets AND being hailed as liberators. (!) So, no, Augustus was not eager to meddle in client Kingdoms politically, at least not on the level of micromanaging local head-counting systems.
Thus, Caesar's 27 BC decision to start provincial registrations is the only event that we know of which could possibly be what Luke was trying to reference. Luke's Greek, of course, refers to something called the "oikoumene" - literally, "the inhabited world", but which some translators believe is more closely rendered by "the civilized world". That second sense, in Roman eyes, could mean that Luke was referring properly to the jurisdiction of all Roman Provinces. When Augustus decreed the provinces should be censused, one could nearly just as well say that Augustus decreed the oikoumene should be censused.
Herod's Judea, Civilized? Well, he was working hard to get there... but no, it wasn't. Not hardly, by Rome's standards.
By the way, Luke 2:1-5 is MOST CLEAR on one point - that a Roman census of some sort took place within Herod's Kingdom while Herod was still in charge. Such a thing was completely unprecedented, so far as we know. So, accepting Luke's clear insistence on this point, the primary historical question must be - how, when and WHY did such a unique census take place?
This raises many more questions, as does the issue of Quirinius. However, if Luke meant for 2:1 to refer to an event which took place in 27 BC, and if Luke 2:2 refers to the well known Governor of the year AD 6, then this odd juxtaposition may be more of a boon than a hindrance. The combination of these references - twentyish years before Jesus' probable birth and twelve or thirteenish years after - make it appear Luke was trying to collapse more than three decades of time into one sentence. In other words, whatever Luke meant to write, in that sentence, he was packing it in very densely, perhaps trusting his readers to sort out that which should have been somewhat familiar to them at the time.
Again, other more pressing historical questions remain; most critically, how and why did Augustus decide to actually go do this thing? But regarding the text itself, and regarding its composition, it seems most likely that Luke was simply trying to reference the two most critical bits of well known Judean tax history - one from 27 BC and one from AD 6, in between which the Lord became human.