For institutional investors, this is a very complicated question. There's always been ways that they are front-runed by more nimble traders.
Currently, the state-of-art for institutional investors are to use buy-side algorithmic tools to slice and dice their orders into 100 lots and obfuscate their huge lot order (> 100,000 shares). Algo's include iceberg which continuously monitors the quotebook throughout the day and execute a portion of customer's orders when the market turns in favor of direction of the order and with an order size not to disrupt the current momentum of the stock's price action; also special darkpools are used at mutual funds such as Liquidnet where no official quotebook are maintained, instead it's more of an instant messenger tool that is limited to only mutual fund managers who chit-chat and close large block deals. Any participants who are suspected of front-running are banned.
As for your original concern, yes spread of a few cents per share is very important for institutional investors when you have to execute a order for 100,000+ shares; and there's actually a whole technology vendor space in finance called transaction cost analysis that fund managers look at to see what the hidden cost of their trades are. As whether HFT actually help institutional investors, it's analogous to saying whether defense contractors help the federal government; some say it's symbiotic relationship, some say it's a parasitic relationship for all and symbiotic for some. Meaning most relationships on Wall Street like on Capital Hill, revolves around wine & dine & shifting allegiances. Quite a few HFT firms have tried to open a new line of revenue by opening up their high-speed infrastructure to institutions to execute their iceberg orders. Yet some of the HFT algorithms' intentions are to sniff out iceberg orders and front-run them.
How do you not love the idea that large institutional investors exploit covert channels to place orders?
As crazy as this sounds, this isn't an HFT thing or a modern markets thing: execution of large block trades is one of the fundamental basic problems in trading. It's just that now we're solving it with a system of dueling robots.
Currently, the state-of-art for institutional investors are to use buy-side algorithmic tools to slice and dice their orders into 100 lots and obfuscate their huge lot order (> 100,000 shares). Algo's include iceberg which continuously monitors the quotebook throughout the day and execute a portion of customer's orders when the market turns in favor of direction of the order and with an order size not to disrupt the current momentum of the stock's price action; also special darkpools are used at mutual funds such as Liquidnet where no official quotebook are maintained, instead it's more of an instant messenger tool that is limited to only mutual fund managers who chit-chat and close large block deals. Any participants who are suspected of front-running are banned.
As for your original concern, yes spread of a few cents per share is very important for institutional investors when you have to execute a order for 100,000+ shares; and there's actually a whole technology vendor space in finance called transaction cost analysis that fund managers look at to see what the hidden cost of their trades are. As whether HFT actually help institutional investors, it's analogous to saying whether defense contractors help the federal government; some say it's symbiotic relationship, some say it's a parasitic relationship for all and symbiotic for some. Meaning most relationships on Wall Street like on Capital Hill, revolves around wine & dine & shifting allegiances. Quite a few HFT firms have tried to open a new line of revenue by opening up their high-speed infrastructure to institutions to execute their iceberg orders. Yet some of the HFT algorithms' intentions are to sniff out iceberg orders and front-run them.