Gross loans and receivables 2017 = $273,188k
Total Provisions = $16,677k
Provisions/GLR = 6.1%
This same ratio was at 5.9% for 2016. As far as actual bad debt write-off goes in the 12 months (i.e. the loans they actually wrote off rather than the total loss value they expect at reporting date on the gross book) then this is the number quoted above as 4.5%. The unsecured part of this business would be highly expected to incur losses in the 6%+ range given the nature of the customer base and this average risk is priced into the arrangements, so provisions/losses in these ranges aren't of any real concern. A large spike or long term negative trend would be. One would expect the provisioning percentage to drop as the move towards more secured lending occurs, but also note that the pricing will be less on this business (you pay more for borrowing on a credit card than a secured personal loan for example).
As Tattie outlines above they will adopt a new accounting standard in the coming year that takes a through cycle loss approach (i.e. expected rather than incurred loss model). This will be expected to increase the overall provisioning percentage on adoption but won't change profitability through the cycle.
I don't see the level of bad debts or provisioning as any concern unless they spike unexpectedly (and they would have to spike quite significantly...). It is this scenario that would be of concern as there wouldn't have been the appropriate pricing for risk and thus average margins on the loan book would get compressed.
While somewhat out of date, as a reference point, look at the following chart from the major banks:
View attachment 932569
You will see unsecured credit card lending with a credit loss rate of >3%. The customer set of MNY is generally of a worse credit rating to that of the big banks and they charge the customer more as a result (i.e. when they price the credit they will include the assumption that more of their customers will default and end up not being paid back). The risk is that they misprice this risk (undercook the risk), not that losses turn up as expected that the average customer pays for.
DYOR