Good news! The conventional wisdom about divorce is a myth. Half of all marriages are not actually ending in divorce. Not by a long shot and not for a long time, according to a smart but frustrating report by Claire Cain Miller for The New York Times's data-driven division, The Upshot.

The piece breathes a sigh of relief. Finally, the bad trends are abating, maybe even reversing for good after a difficult period of adjustment to the sexual revolution, which taught us once and for all that marriage is for love. This happy state of affairs stems almost entirely from the great rectitude and pragmatism of liberal cultural values to boot. We're richer, more liberal, and, well, just better. We love each other more than our grandparents did. Data says so.

Nope.

It turns out that this bit of good news from the Times' hard-nosed ledger sniffers turns out to be a Styles section trend piece in disguise. Let's start with the big sell of the article, the assertion that "marriages in this country are stronger today than they have been in a long time." But that's actually a rather limited observation. Yes, the marriages that do happen do not break up as quickly or as often as marriages from 30 years ago. But the truth is that family instability continues to worsen in the United States. As David Frum pointed out, a declining divorce rate is perfectly consistent with an ever-falling rate of marriage and a rising rate of out-of-wedlock childbirth.

If there is one finding in social science rock solid enough not to be mere post-propter-hoc speculation or a bias toward correlation, it is that children do best in intact, two–biological parent homes. Which means that, despite the happy news for Times readers, children as a class are having a rougher time as marriage culture continues to erode for the rest of us.

Then there are problems with the details of the article. The Upshot writes that 5 percent more marriages from the 1990s are reaching their 15th anniversary intact than marriages from the 1980s and 70s. Alas, that stat says too little when you take into account the larger story of marriage's decline. In 1956, the rate of out-of-wedlock births among non-Hispanic whites was 1.9 percent. In 1990 it was 16.9 percent. Now it is around 30 percent. The rate of out-of-wedlock births among blacks is much worse, a telling indicator of where those with lower incomes are headed.

And, as a side note, even a 15th anniversary is no guarantee of lifelong commitment. See the rise of elder divorce.

What of The Upshot's claim (echoed by others) that marriage has become more successful now that love, thanks to the sexual revolution, has become the ultimate criteria for marriage? The idea is that, once for all, people have put aside concerns about property, dowries, and estates and embraced companionate marriage.

This claim is risible. The idea that most marriages were previously loveless, or at least stitched together with much greater indifference to love, is a self-flattering fantasy that recurs constantly. As Ferdinand Mount recounts in his book The Subversive Family, historians since the Victorian era have variously pointed to the medieval period, the Renaissance, or even the Tudor era as a time when romantic love and marriage for love were finally discovered or invented. In fact, testimony to love-marriage has been recorded in every era. Our notions of history are corrupted by the fact that the lives of royalty and other aristocratic classes, where arranged marriage was more common, were better recorded.

Speaking of which, our idea that property and prospect matter so much less now is silly. Marriage patterns are becoming more narrowly class-based than before. There were more marriages contracted across socioeconomic chasms in the bad old 1950s and 60s. And while we like to think that marriages of convenience are a thing of the past, we all know that they exist today, and that the forces of romantic love on the one hand and economic interest on the other can wax and wane. A romantic marriage can easily devolve into a business.

The Upshot also gives fulsome credit to progressive advancements of cohabitation and later-marriage for the slight drop in divorce among the (shrinking) married population. But these are more common on lower economic rungs where divorce hasn't declined as much, or where marriage doesn't even occur. Cohabitation is said to be helping marriages at the top by allowing bad relationships to disintegrate before divorce is necessary. But looked at from a wide-angle perspective, cohabitation looks like a substitute for marriage for many others.

And what The Upshot doesn't consider is whether inequality itself is helping the marriages of the upwardly mobile. The data shows that people who already succeed in many aspects of their life are making successes of their marriages. Far from a progressive dream, we may be returning to the two worlds of aristocracy. A married upper class and an unmarried peasantry is exactly what you see when you look at the British Isles in the early 20th century. Those living in converted Abbeys could keep their marriages together, but 65 percent of Ireland's population was unmarried at the same time, the highest portion in the Western world of that era. There's just more incentive to hold together the "estate of marriage" when the married couple have property that might qualify as an estate.

It's a downer, I know. But far from a trendline of unqualified marital bliss, the prospects for marriage look bleak. And the improved prospects for a certain class of married person may not be caused by liberal values at all, but may be a side effect of concentrated inequality.

The real trend is that marriage is for richer, not poorer. And our only proximate hope is that the rest of America will try to imitate the slightly better marriage patterns of the rich and famous.