Elevation

So I'd argue there's actually some pretty big differences between paying up front and eating low, first year pay. It all comes down to risk and incentives.

"Normal" practice for an airline to screen candidates based on whatever they need. Since they are eating the costs of training the airline or charter company needs to be pretty selective of their candidate. There's real money riding on whether this person is going to make it through training. Of course other factors are at play, but since your chances of making it through training are tied to the airline's cash, your knowledge and skill count for something. The airline has an incentive to select aviators at least in part for their ability to aviate. Airlines make sure their investments are secure. The people around us are safer for this.

PFT has the candidate ante up a large amount of money. The airline has much less invested in whether you make it through training or not. Now those other factors become much more important. Not only do your looks, connections, etc. matter, but your ability to pay for this training also affects whether you will or won't be selected. In this model the airline has much less incentive to select based on ability since the cost of a pilot is low. The people around us are less safe due to this.

So when we look at Atlas' model we approached a PFT-like result without having people pay up front for training. Less money was tied up in our ability to aviate than our ability to fill seats. We fell into a similar trap, only instead of selecting for looks and connections, we selected for whoever would sit down, shut-up and not make trouble. While the airline was investing money in training, the cost of our pilots was low. That meant other incentives prevailed. Now people are dead.