Many managers look at the cost differential of hiring a high-end designer versus getting their back end dev to mock something up, and ask themselves "Will I get another $X,XXX amount of business in return for this investment?" Depending on their personality, they will then answer "yes" or "no", and act accordingly.
The reality is the manager is a very bad person to make that call - as they are intimately familiar with the business, and therefore generally very bad at judging how a person who is encountering the business for a first time will respond to a design.
Testing the new design is expensive though, as it requires an outlay of $X,XXX without any guarantee return of results; therefore the manager takes the conservative approach and hires the cheaper designer, as at least that way she/he won't have the issue of having spent money and not being able to show any results from the investment.