I applaud Steve's reasoning and conclusions in this well written argument, especially his last sentence, "that utilities, regulators and customers should have a strong interest in wringing efficiencies from the grid itself first rather than just hectoring customers to change when and how they use electricity."
Historically, regulators have approved capital investments into rate base to ensure system reliability, with the understanding that over the long term, those investments would be paid for in rates by a captive ratepaying public that approximated society at large. In other words, investing in the grid owned by utilities was a societal good because it ensured a reliable electric infrastructure. Because the grid was virtually the sole supplier of electricity back in the day, the grid was built with tremendous reserve capacity so that it could serve the dramatic increase in system load on critical peak days (e.g., a few late afternoons in September in much of the US). That strategy worked quite well, as Steve attests, but it has been an expensive approach to keep investing in more supply side resources, rather than more efficiency.
The question of demand side solutions has been raised for some time, but rarely have such options received serious consideration because the entire ecosystem was oriented around supply side solutions - more generators, more wires, more fuel. As for eliminating waste, requiring grid operators to get more out of their systems with more efficient operations is difficult and unlikely to result in large benefits for the shareholders at any rate. In contrast, approving advanced meter investment has merits and promise for a demand side focus, finally, and it is a solution that also works for shareholders, and especially makes sense when the federal government has subsidized significant costs of this route. But the promise of AMI to reduce peak hinges on multiple additional steps to meet its goal of changed user behavior and peak shifting, making it a higher risk prospect to acheive its stated goals than appears at first glance.
In contrast, what if distribution automation had gone first, requiring zero cooperation from consumers? Such grid modernization would attack those operational inefficiencies directly in control of engineers and operators, systematically driving out waste and improving system reliability. And direct solutions with fewer contingencies have a higher probabilty of success. But there's more to it than that. Having wrung every efficiency they could out of grid optimization, utilities could go back to the regulators with their next plan, to move on to collaboration with system users to elimiate wasteful behavior on the demand side. And they would have the advantage of having saved customers money by going first on their supply side, thereby deferring rate increases, and thus making a stronger argument to move on to the next logical step: AMI, MDM, digital billing, TOU rates, DR programs, leading to changed user behavior and peak shifting.
I personally believe we need all of the above - grid optimization, system effciencies, greater reliabilty, improved security, and peak shifting to avoid additional capital expenses - but the questions in considering these options should include Which step is most effective? Which has the highest probability of success? and Which has the highest ROI? - thereby providing the most bang for the buck with the least disruption.