"Likely" because for the scenarios/markets where international payments work poorly there's no "actual numbers" in this situation, the pricing is highly dependent on circumstances, the specific institutions involved, the currencies involved (there may be multiple, and for some countries a currency conversion may be mandatory), the payment scheme involved (depending on the country/bank pair there likely will be multiple payment channels with different price/speed/reliability tradeoffs) and who is paying (which affects a lot the prices you may get). For such deals the sending institution will generally refuse to quote a fixed generic price, because the price will be different depending on all kinds of factors that, in some cases, will be revealed only after the payment is delivered or returned.
An bank payment from Nigeria to Peru might reasonably cost $20-$100, and I wouldn't be that surprised if in a particular situation it would come out to $10 or $300 (like, making a typo in some field of a non-validated foreign payment, or having an unclear tax situation on the goods for which the payment is held - that can easily cause e.g. a $100 extra unexpected processing fee for the payment). Also, you may get amount-proportional fees, and a 0.5% fee can potentially be huge for larger payments. On the other hand of the comparison there is also significant variability for the extra costs of the cryptocurrency payment - as the original question stated, including the cost and exchange rate variation of buying crypto for fiat in the sending country and selling crypto in the receiving country.
So it's likely but not certain that it's more expensive than an equivalent crypto payment + the to-crypto and from-crypto conversion, it depends on unknowable circumstances, and the variability and unpredictability of such fees is a big drawback on those types of international payments - unlike e.g. the intra-EU case where the fees would be predictable and low - perhaps zero, depending on your bank.