If the festival's date is variable, the date it commemorates is far more so.
Two chronologies have long competed. The long chronology, from the Sri Lankan Pali sources, places the death in 544 or 486 BCE depending on how the regnal figures are handled, giving a life of about 563 to 483 BCE. That is the date in older reference works and in the Buddhist Era used officially in Thailand, Sri Lanka, Myanmar and Cambodia, which counts from 544 BCE and is why those countries number the year in the twenty-five hundreds.
The short chronology, from Indian and Chinese sources, places the death around 368 BCE.
Since a symposium organised by Heinz Bechert in the 1980s the scholarly centre of gravity has moved to a date of death somewhere between roughly 410 and 370 BCE, about a century later than the traditional figure and closer to the short chronology. The anchor for all of these is the reign of Ashoka, independently datable because his inscriptions name Hellenistic kings whose dates are known.
Excavation at Lumbini published in 2013 reported a timber structure beneath the later brick shrine with radiocarbon dates in the sixth century BCE, which the excavators presented as supporting the earlier chronology. The interpretation of the structure has been questioned.
The practical result is that the official Buddhist Era year, printed on documents in several countries, rests on a chronology most specialists no longer accept. This causes nobody any difficulty, which is itself worth noticing.