The University of Michigan and Thomson Reuters are getting a lot of heat from recent reports about a million-dollar deal whereby Michigan releases a data feed with survey results a full two seconds before the results are available to select Thomson clients—who in turn get it five minutes before the public. I do not know enough about this deal to comment generally (though I am at umich.edu I am far removed from anything having to do with the consumer survey). However, I can say something about the connection to high-frequency trading.
The “ultra-low latency distribution platform” obviously caters to algorithmic traders, and there is evidence that these HFTs are generating plenty of activity on data release. However, payment for early access to information seems to me different in kind from some of the other extreme measures HFTs take to gain information advances.
Paul Krugman calls this an example of “real resources being devoted to the socially useless task of getting an economic number slightly before the hoi polloi”. I agree with the “socially useless” part of that, but don’t see much in the way of “real resources” being used here. There is little extra expense entailed for Michigan to give information to some parties at 10:00:00, others at 10:45:00, and yet others at 10:44:58. Just whatever (presumably quite small) effort was required to set up the special distribution platform. This contrasts starkly with the major resources (putting in new communication lines, special-purpose computing hardware, etc.) employed in the latency arms race generally to shave microseconds off response time.