Hey @booster, I have to question your maths a little on this one.
If we don't have issues with SBD conversion, supply of Steem is expected to increase by about 8.5% this year. For holding Steem Power, we only earn about 1% - 1.5%. Witnesses get a little under 1%. So that leaves us without about 6%. There are several large accounts such as @steemit, @mottler, etc that do not vote at all. Most other accounts do not vote efficiently. Bots such as booster maximise SP by voting 10 times a day every 2.4 hours. So the 6% could become 18% or even higher for bots because of this voting efficiency.
If someone is just delegating to a bot. The 18% would be less as bots get the curation reward and some is given out as ROI to vote buyers. So 18% would probably look more like 13%. That would give us vote selling revenue + interest = 14%. If we assume a constant market cap, we are left with 14 - 8.5 = 5.5%.
As more people delegate to bots, this 18% will start to fall as more SP is being used efficiently. I think at the moment about 15% (including smartmarket and minnowbooster) of SP is going into vote selling. If this doubled to 30%, 18% could drop to about 15%. If it went up to 50%, it could drop to 10%.
The increase in the number DApps also means that Steemit Inc SP is becoming active. This active SP will lower the rewards from vote selling.
The best way to get a great return is to increase market cap. For example by marketing Steem on other social media, getting Steem on more exchanges, giving Steem DApps more exposure. People such as @surfermarly, @nathanmars, @therealwolf, and @theycallmedan just to mention a few have made great strides in this area.
Another great return could be revenue sharing from advertising on DApps. I am sure it won't be long before a DApp offers delegators a cut of their advertising revenue. I don't think we need to limit ourselves to just earning from the existing rewards pool.