The efficiency argument implicitly assumes that electricity will be scarce, but this is unlikely to be the case if the majority of the world's energy needs are

The efficiency argument implicitly assumes that electricity will be scarce, but this is unlikely to be the case if the majority of the world's energy needs are met through wind and solar. To give you an idea of why, Germany and California, at less than 50% renewable penetration, already curtail around 9TWh of energy a year due to mismatched supply and demand. This is enough to charge on the order of 130 million BEVs. As renewable penetration increases, so too does the need to provide statistical confidence in generation capacity. If you want to ensure that your wind farm is able to provide X units of power, say, 80% of the time, you'll be overproducing pretty much constantly. Connecting cities doesn't really change this either, because the problem is one of stochasticity and confidence. In order for two sites A and B to cover shortfalls for one another, it has to be true that both A and B are capable of overproducing, and that at least some of the time, both will overproduce (unless they're perfectly negatively correlated). If you want to ensure that A and B both have enough power with some statistical confidence, then your system will produce excess most of the time.