When do other banks start to publish SPF records?

In the past a lot of phishing was going towards customers of the Dutch bank Postbank. It continued for years and when the bank finally merged with ING the phishing attacks adopted the new name quickly. In both cases the bank was publishing closed SPF resource records in DNS so third party systems could determine of an e-mail really came from Postbank or ING. And with a few rules for SpamAssassin for example most of the phishing can be stopped.

The last months phishing attacks for both Rabobank and ABN Amro increased a lot. Most phishing e-mails from Rabobank are being caught by the bayesian filter for now, but for ABN Amro aren’t always detected. This makes me wonder why those banks don’t publish SPF resource records in DNS? Is it really that difficult? Or is the cost for fraude smaller, then for a denied e-mail?