What do you mean by this?
Latency arbitrage only exists because of the physical distances between the multiple exchanges.
Example :
Your looking to buy 1000 shares of stock ABC at 10.00$ and you can see on the order book that their are 1000 shares available with an ask of 10.00$
500 of those shares are at exchange XYZ and 500 shares at exchange UVW. You send your marketable limit order or market order to buy 1000 shares at the current offer of 1000 shares at 10.00. Unless your broker has sophisticated enough execution programs to make sure that order which is for 500 at one exchange and 500 at the other hit the exchanges at exactly the same moment, half your order is going to get picked up by a HFT trader.
They seeyou pick up 500 at one exchange and race off to the other exchanges to pick up any offers and sell them back to you at lets say 10.01 or 10.02 or whatever the spreads allow them to be the best in the book now. If the offer was 10.00 and the 2nd best offer was 10.25, they can pick those 500 at the second exchange at 10.00 and quickly flip them to you at 10.24 beating the rest of the book. Bit of an extreme example but is possible.
Payment for order flow really only exists because a lot of institutions can't compete on the open book against HFT and they want pieces of the "dumb money" (retail) so their willing to give us price improvements for our order to get a piece of the action.
The idea of just IPO your own company isn't exactly a great idea either. We need IPO's to keep the market going but we also need quality IPO's. For example take a look at the TSX venture which is just an absolute ghost town. Way too many garbage companies eating up what little liquidity is available there. Could even really look at the OTC markets.
With a quick screen of U.S stocks I see just shy of 8000 stocks listed. If I add a filter of a 90 DSMA of volume > 1million and a stock price > 5.00$, that world drops to 1000 stocks.
Traders want liquidity, it keeps transaction cost down and creates more oppurtunities. Liquidity comes greatly from HFT and for a market to offer a good depth of book, you need participants. The participants all congregate where everyone else is. You can have 1000's more of IPO's but you need participants to make the market.
Now don't get me wrong. I think a lot of HFT actually offers some usual service to the markets. For instance take someone running a strategy trading inter-listed stocks and their corresponding currencies. Theirs an arbitrage there that they will dominate and this helps keep the markets locked step. Perhaps someone running a statistical arbitrage between a cash index and its corresponding futures market listing.
I really think it should all be ran on a central exchange with a central order book. However then this creates the oppurtunity for monopoly for an exchange which is also never a good idea. It's kind of this fine line.
Here's a couple good podcast that kind of go over these topics.
https://chatwithtraders.com/ep-088-dave-lauer/
https://chatwithtraders.com/ep-075-dan-aisen-iex-trading/
https://chatwithtraders.com/ep-071-eric-hunsader/
Also these little documentaries are well put together.