Joel Waldfogel is well known for having pioneered the economic case against gift exchange, arguing basically that it causes enormous amounts of dead-weight loss and unnecessary search costs. Everyone could just keep their own money and buy what they want. The basic microeconomic logic of this seems unassailable to me, and yet in an Initiative on Global Markets survey of leading economists surprisingly few were willing to come out against Christmas. The pro-Christmas arguments they offered were surprisingly weak—basically amounting to asserting that the warm fuzzies some people get from some gifts somehow obviously outweigh the stress and anxiety that other people experience during the gift-giving process.
The real economic case for Christmas is macroeconomic. There's an old Keynesian saying: It takes a lot of Harberger triangles to fill an Okun Gap. Which is to say that the total amount of harm done by micro-inefficiencies is small compared with the massive harms associated with the macroeconomic slack of recessions.
Most people do not realize this because the media generally report on seasonally adjusted macroeconomic data, but there are gigantic economic fluctuations associated with the passage of the seasons. As Robert Barsky and Jeffrey Miron have shown, the seasonal business cycle is comparable in size to the official business cycle. The economy in effect enters a recession every January, grows steadily over the next several months, slumps again in the third quarter, and then booms in the fourth quarter. Anton Braun and Charles Evans have shown a couple of further surprising facts about this seasonable business cycle (PDF). One is that not only does total output boom in the fourth quarter, but total-factor productivity booms as well. TFP is the X-factor in economics often labeled "technology" but in fact simply meaning whatever element of productivity isn't accounted for by the use of more labor or more capital. The TFP boom is moderately surprising since there's a boom in Christmas temp hiring (and then big layoffs in January) meaning the marginal workers in Q4 should be lower quality than the marginal worker at other times. The second surprising fact is that they show the entire seasonal cycle is essentially just driven by Christmas.
Put those together and it looks like Christmas not only leads to more people working, but faced with a surge of demand, managers somehow manage to get everyone to work smarter and more efficiently even as the total number of workers grows.
Tyler Cowen rightly posits that this boom and bust cycle isn't exactly efficient, but it seems to me that the inefficient part is the bust not the boom. We experience an annual January-March recession that would be considered intolerable if not for the fact that the Bureau of Labor Statistics seasonal adjustment algorithm makes it go away. But we should be aspiring to have Christmas-esque levels of production all the time. The fact that some of the Christmas output is wasteful (presents people don't want, etc.) isn't a big deal compared with the fact that there's much more output overall.
A number of people have responded to me that perpetual Christmas would entail a lower savings rate and a higher consumption share of the economy, but I don't think that's the case. My Christmas shopping and holiday tips are other people's income, so consumption can rise without the consumption share rising. Christmas-related income booms are going to be very heterogeneous (hair stylists get tips, retail clerks get extra shifts, but lawyers and investment bankers just take some time off) so modeling this properly would be fairly challenging, but it's at least not obvious that you'd be talking about dissaving.