Competition for the wallet with lowest fees doesn't happen until there's a large user base that's in for the long haul. Even then, many users will just use the setup they're used to unless the fees are ridiculous.
Ask bytemaster if he's in for the long haul or not. Shortsighted decision making can leave businesses crippled in the long run. Ask yourself, if there were a perfect replication of BTS, call it BTS2, and if BTS has 3rd party referral incentives and BTS2 has protocol level referral incentives, which DAC, BTS or BTS2, performs better over the long haul, and even short haul for that matter.
Adding it at the protocol level is just unnecessary, and adds significant long term privacy concerns.
I may see where you're going with this, but public reputation is where it's going to be at!
This is a solved problem. The model used by the light and web wallet developers will also function for referrals.
The dev team has made no indication that they feel this is a better solution, and I gather that they see the protocol as a better base for marketing incentives.
In summary, you made the point yourself,
3rd party referral incentives is not a solution for the long haul because competition will drive down fees. BTS may face less competition than wallets, and therefore its products (including their fees) have greater inelasticity, which means referral marketing protocol level is better suited for the long haul.