Effectively the entire corpus of around a thousand pendants falls in the ninth to eleventh centuries. That is also the conversion period, so the distribution is compatible both with an old practice becoming archaeologically visible and with a practice that begins as a response to Christian cross-wearing. No securely dated earlier example has been published, and no positive account of an earlier phase exists, so this is an evidential gap rather than a contest between established positions.
What would resolve this A hammer pendant, or a mould for one, recovered from a securely dated context earlier than the ninth century, or a systematic re-examination and re-dating of the existing corpus establishing that no such context exists.
OpenViking Age and Migration Period archaeology, with finds dating and metalwork typologyJohn Lindow, Norse Mythology: A Guide to the Gods, Heroes, Rituals, and Beliefs