The Mysterious Bah Humbug Toward Christmas Shopping

November US retail sales were a disappointment. Black Friday didn’t draw big crowds. Christmas spending intentions as tracked by Gallup fell from October to November and are down almost 9% from last year after adjusting for inflation. Retailers that benefit from Christmas shopping (i.e. not gas stations, supermarkets, building supply stores and auto dealers) increased employment by only 2.7% in November, the smallest such buildup on record.

Get the full experience.Choose your plan

Inflation and fears of recession account for some of this, with the shift to online shopping playing a supporting role in reducing in-store crowds and shifting some of the holiday retail employment buildup to warehouses and delivery services. But something bigger may be afoot, given that Christmas shopping — by which I also mean shopping for Hanukkah, Kwanzaa and any other end-of-year gift-giving holiday — appears to have been declining for decades relative to shopping during the rest of the year.

I looked up these statistics because I write at least one holiday-season column every December (last year’s was about Christmas poet Clement Clarke Moore) and was casting about for a topic when a colleague suggested examining the economic health of Christmas. Using December’s share of annual retail sales as the metric was my idea, but the list of seasonal retailers is derived from a 2019 Bureau of Labor Statistics examination of the declining holiday hiring surge.(1) And no, December’s decline isn’t just about people shopping earlier — including November doesn’t change the picture much.

The Census Bureau’s fully comparable, fully revised numbers are available only back to 1992. But the agency started measuring retail sales decades before, and after some digging through old press releases, back issues of the Bureau of Economic Analysis’s Survey of Current Business and a 1997 Census publication with revised data for the previous decade, I was able to produce this:

There are clearly some comparability issues between the old data and the new, but the numbers do give the overall impression that December’s share of retail sales was steady to rising before 1990, with dips during recessions, and has been falling ever since, with last December’s 11.4% share the lowest on record. Last December was admittedly an anomaly, with the omicron wave of Covid-19 persuading many people to stay home for weeks. But the December 2018 share was just 11.6%.(2)This seemed like a topic for Joel Waldfogel. As an economics professor at Yale in the early 1990s, Waldfogel surveyed students about how much they valued the holiday gifts they received and how much they thought the gifts had cost and came up with an estimate that between one-tenth and one-third of the spending on those gifts was “deadweight loss” — economic value that the givers sacrificed but the recipients didn’t gain. His findings were published in the December 1993 issue of the American Economic Review under the title “The Deadweight Loss of Christmas” and, well, look at what happened:

When I jokingly emailed a version of this chart to Waldfogel, now a professor and associate dean at the University of Minnesota’s Carlson School of Management, I received an interesting response. A few months ago, he, too, had decided to look at what was happening with Christmas shopping, using a metric similar to but simpler than mine — December’s retail sales minus November’s. He had run these numbers a decade or so ago and hadn’t seen much of a trend, but now he clearly did (it was similar if less pronounced when he subtracted December’s retail sales from October’s or September’s). Here’s what it looks like using the same seasonal retail sectors as in my other charts after adjusting for population and inflation.

Waldfogel also looked at pre-1992 US retail sales, including partial statistics available all the way back to 1914, and compared them with changes in incomes. As he details in a short paper titled “Holiday gift giving in retreat,” to be published soon in the journal Economics Letters, real holiday spending rose faster than incomes early in the 20th century and at about the same pace from 1935 onward, but since 2000 “gift giving falls as disposable income continues to rise.” (My fellow Bloomberg Opinion columnist Tyler Cowen noticed the paper, too.)

This would seem to imply that holiday gifts have become an “inferior good,” a product — such as rice in Asia or beer in Germany — that people consume less of as their incomes rise. This doesn’t square with Gallup survey data showing that higher-income respondents still have higher Christmas-spending intentions, but Waldfogel said he still wonders whether perhaps “as we’ve continued to get richer we don’t need this treat of surprise. We just buy things for ourselves whenever we need them.”

The Gallup surveys definitely back up the notion that Christmas spending is down when adjusted for inflation, although in this case the decline has occurred not steadily but in two big drops, in the early 2000s and in 2008.

Recessions help explain both those declines, and it’s possible that the economic dislocations and rising inequality of recent decades have played a role in driving down Christmas spending. Something else of importance that happened in the 2000s is that broadband internet became widely available. This pretty quickly killed off one popular Christmas present, the CD, and the same happened a little later for DVDs. Even with the recent modest resurgence of vinyl, music and movies are now paid for chiefly by subscription and don’t show up as Christmas spending. It’s not that Christmas itself is on the outs — my current Spotify playlists are dominated by Christmas songs, and more Christmas movies are being produced than ever before — but the consumption of all that holiday cheer is invisible in the retail sales data. Along somewhat similar lines, Waldfogel speculates that there’s been a rise in “experience gifts” such as trips where the spending takes place in some month other than December.

The aging of the population is another possible factor, with the percentage of Americans who are younger than 18 — the prime gift recipients, one would think — falling steadily since the late 1990s. The younger-than-18 share fell even faster in the 1970s, and Waldfogel’s results didn’t change when he redid his analysis using the younger-than-20 population rather than the overall population. But the rise in the 65-and-older population share from 12.4% in 2005 to 16.8% in 2021 is unprecedented, and one would imagine that senior citizens are especially unlikely to want more stuff.

What about the impact of the “The Deadweight Loss of Christmas”? The most prominent American critique of Christmas commercialization is probably “A Charlie Brown Christmas,” which first aired on national television in December 1965 and has been rerun every December since.(3) It does not appear to have turned the tide in its first quarter century, so you wouldn’t think an academic paper could, and Waldfogel certainly doesn’t claim that his has (“it would be hard for me to confidently draw that conclusion from the available evidence,” he says). But the dissemination of his arguments — in a 2009 book, Scroogenomics: Why You Shouldn’t Buy Presents for the Holidays, a growing economic literature and lots and lots and lots of holiday media coverage — has at least coincided with the rise of an approach to gift shopping that is remarkably compatible with them.

Waldfogel does not oppose giving gifts, he just warns that buying things for people whose habits and needs you’re not very familiar with (nephews, nieces and grandchildren, say) can be pretty inefficient. “We are able to choose gifts well only for people we know really well,” he wrote in 2015. “With everyone else, we might be better off giving cash or gift cards.” Americans have been doing the latter in an ever bigger way, with ResearchAndMarkets.com forecasting $189 billion in gift-card spending this year — which amounts to 70% of last December’s seasonal retail spending and is about four times higher, adjusted for inflation, than estimated gift card spending in 2001.

The Census Bureau counts this spending as retail sales only when the gift card is used to buy something, which has the impact of dispersing Christmas gift spending over the course of the year. Gift cards can also sometimes put off that spending forever, with CreditCards.com estimating in August that Americans currently hold $21 billion in unused gift cards and store vouchers — and that’s based on a poll, so it doesn’t count all the cards people have forgotten about.

Still, none of that should have an impact on the answers people give to Gallup’s Christmas-spending-intentions questions, and gift cards thus cannot fully explain the spending decline. It remains a Christmas mystery. Or, as Waldfogel puts it, “I think I’ve raised as many questions as I’ve answered.”