We also have to consider how spammers will optimize their operations to get around this limit.
There is little cost to adding spam to the blockchain.
True. At the moment, it's cost free to create accounts via steam, and spamming allows a spammer to earn either through random upvotes given by users, abuse of upvote bots, or circle upvotes - in fact, the more accounts they create, the higher the rewards because each Steemit-created account has a total active SP of at least 15.
In other words, it's a cost-free venture (disregarding server and time costs for running a large ring of accounts), with the chance of reaping rewards. Profit is guaranted to be non-negative.
But will raising the threshold impose a cost?
Not really. There's no direct monetary cost imposed, and only an "economic cost" from less expected profit. The outcome will be less expected profit per account and spam post/comment, but it's still non-zero. The time cost for a true spammer who uses computer programs is close to 0 once there is some system set down.
If I were a spammer, I would get around this by creating more accounts to put out more posts so I hit my target return, reap more self-upvotes to get over the limit, or simply use upvote bots to hit the minimum required.