Did the New Mass Empty the Pews?

Traditional Catholicism · Evidence Review

The claim is made constantly on both sides, and almost always without numbers. Here are the numbers, including a 2025 econometric study that changes the picture, three natural experiments most traditionalists have never had to answer, and the one argument against the reform that needs no statistics at all.

Ask whether the Novus Ordo emptied the pews and you will get two confident answers, neither of which is usually accompanied by evidence. The traditionalist says the collapse is self-evident: look at the seminaries, look at the convents, look at Sunday. The defender of the reform says the decline was already underway, that Africa is exploding, that correlation is not causation.

Both are partly right, and each is defending his weakest ground. The purpose of this article is to put the actual data on the table: from the Vatican’s own statistical yearbook, from CARA at Georgetown, from Gallup’s postwar survey series, and from a major 2025 econometric study that has changed what a careful person can say about this.

The conclusion, stated up front so you can argue with it as you read: the evidence supports a real, sharp, and specifically Catholic break in the 1960s that secularization alone cannot explain. That is a genuine finding and it favors the traditionalist. But almost nothing in the evidence points to the Missal as the mechanism, and three natural experiments make liturgical causation very difficult to sustain. The strongest argument against the reform is not the one traditionalists usually make.

Most of this debate is unresolvable because the parties are arguing about different propositions at once. Separate them and each becomes answerable, and note that the last of them, which is the strongest claim against the reform, has nothing to do with attendance at all. Here is where the evidence lands; the rest of the article shows the work.

Start where the traditionalist is strongest. The Official Catholic Directory has counted American Catholic institutional life every year since long before the Council, and CARA compiles it. The figures are not ambiguous.

The chart below indexes each measure to its 1965 level. A line at 100 means “unchanged since the Council closed.” What it shows is a Church whose membership grew by half while every measure of its sacramental and consecrated life fell off a cliff.

A detail worth pausing on, because it complicates the simple story: the priesthood did not collapse on schedule. In 1975, five years after the Novus Ordo took effect in American parishes, there were 58,909 priests against 59,426 in 1965. Ninety-nine percent. The priestly decline is real and severe, but it arrives in the late 1980s and 1990s, twenty years after the Missal changed. Whatever drove it, it was not working through the liturgy at any recognizable speed.

The sisters, by contrast, fell immediately and catastrophically, down 24% by 1975 alone. That is a different curve with a different cause, and economists have studied it directly: the post-conciliar renewal of religious life radically altered the cost-benefit structure of the vowed vocation, and the orders that liberalized fastest emptied fastest.

Mass attendance is the measure the argument really turns on. Gallup began asking in the 1950s; CARA has run a continuous series since 1970. Together they give us seventy years.

That 1970–75 drop is the traditionalist’s single best datum, and it should be conceded without hedging: the sharpest five-year fall in the entire seventy-year series is precisely the interval in which the Novus Ordo was introduced. Anyone who waves that away is not arguing honestly.

But look at what else occupies that interval. Humanae Vitae was published in July 1968, and the mass public dissent that followed it (priests, theologians, and entire episcopal conferences publicly qualifying a papal teaching) was the single most visible authority crisis of the century. Andrew Greeley, who spent his career on this data, was blunt about which mattered more: it was the birth-control collision, he wrote, “more than any other” development, that “shattered the authority structure.” The Missal and the encyclical hit the same five years. No time series can separate them by inspection.

Here the traditionalist case, as commonly stated, simply breaks. The assertion that the Church is “smaller now than it has ever been” is not a matter of interpretation; it is contradicted by the Vatican’s own count.

The obvious reply is that the world grew too — that this is just demography. It is a fair objection, and the answer to it is the more remarkable statistic:

Across the entire post-conciliar period, the Catholic Church held its share of the human race essentially constant. Not a collapse, not a triumph: a flat line, through fifty years that included the loss of Europe and the winning of Africa.

Global ordinations tell the same awkward story. Worldwide diocesan ordinations were 4,622 in 1970. By 2000 they were 6,814, a 47% increase across the three decades when, on the simple thesis, the new Missal should have been strangling vocations. They have since eased to 5,278 in 2020, still well above the 1970 figure.

This is the standard defense of the reform, and it is asserted far more confidently than the evidence permits. It depends entirely on which country you name.

Where it is true: France was decades into dechristianization before the Council. The sociologists Gabriel Le Bras and Fernand Boulard had mapped whole dechristianized regions in the 1930s and 1940s; the premise of La France, pays de mission? (1943) — the book that launched the worker-priest movement — was that the Church had already lost the industrial working class. Quebec’s Quiet Revolution began in 1960 and gutted practice and vocations before a word of the liturgy changed.

Where it is false: the United States, where sisters, seminarians, and parish-connected Catholics all peaked in or after 1965. And spectacularly the Netherlands, which was at roughly 64% weekly attendance as late as 1966 and then fell further and faster than any Catholic country in Europe.

There is also a dating problem that quietly does a great deal of work in this argument, and it deserves to be said plainly.

Until recently, the honest position was that the causal question could not be settled, because everything declined at once. Mainline Protestantism fell on nearly the same curve with no Missal to blame, which looked like decisive evidence that a general cultural force was at work.

In July 2025 that changed. Robert Barro (Harvard), Laurence Iannaccone (Chapman), and Edgard Dewitte (Oxford) published Looking Backward: Long-Term Religious Service Attendance in 66 Countries as NBER Working Paper 34060. Using retrospective questions from the International Social Survey Program, they reconstructed attendance rates back to the 1920s across 66 countries (over 200,000 responses), and ran a formal event study on Vatican II.

The design is the thing that matters. They compare predominantly Catholic countries against non-Catholic ones, and separately against Protestant and Orthodox countries, defining “Catholic” by a country’s 1900 adherence share so the classification cannot be contaminated by the Council’s own effects. Their findings:

• A decline beginning immediately after the Council, significant at the 1% level and sizable in every subsequent decade.
• Roughly four percentage points of relative shortfall in every post-conciliar decade — about 10% of the period’s attendance mean.
• No pre-trends. Before the Council, Catholic countries were not on a significantly different path from everyone else.
• The effect survives comparison to other Christian countries, which means it is not generic secularization. It is Catholic-specific.

That last point is the one that should change minds. The reflexive answer (“the Protestants declined too, so it was the culture”) is now empirically inadequate. Something happened to Catholics, as Catholics, that did not happen to their Protestant neighbors in the same decades. The traditionalist has been saying this for sixty years and now has a peer-reviewed instrument saying it too.

The honest reading of the study, however, requires four caveats that its enthusiastic circulation on traditionalist sites has tended to drop:

• The event is the Council, not the Missal. The authors date the shock to 1965 and structure their data in decades, so the first post-period runs 1965–1974. The Council’s close and the new Missal sit inside the same bin. The design cannot separate them, and the authors never claim it can.
• The mechanism the authors propose is not liturgical. They lean on Iannaccone’s club-good model (that religious communities derive commitment from cost and distinctiveness, so lowering the barriers reduces participation), and on Greeley’s argument that the Council shattered the perception of an immovable, truth-holding Church. Liturgy appears in their account as one of five changes, and not the one they weight.
• Treatment is at the country level. “Catholic countries” means those over 50% Catholic in 1900, France, Italy, Spain, Ireland, Poland, Latin America. The United States, the Netherlands, Germany, and Britain fall outside the treatment group. The study does not measure the American case at all.
• It measures relative, not absolute, decline. “No pre-trend” means Catholic countries were not diverging from others before 1965. Both groups could have been drifting downward together, and in Western Europe, both were.

So the study establishes something important and narrower than the headlines: the Council marks a real, Catholic-specific inflection. It does not identify which of the Council’s consequences did the work.

Gallup’s postwar surveys break attendance out by age, and the resulting table is the most illuminating single piece of evidence in this whole debate, because it shows both sides something they would rather not see.

Two things jump out.

For the reform’s defenders: Catholic attendance among 21–29 year-olds had already fallen from 73% to 56% by 1965–66, before the new Missal existed. Whatever began this, it began during the Council, not after the Missal.

For the traditionalists: look across at the Protestants. Their young fell from 40% to 31% in the same years, and then recovered, to 36% by the mid-1980s and 39% by 2005–08. Catholic 21–29 attendance went to 35% by 1975 and to 25% by the mid-1990s and never came back. The 1960s hit both. Only one of them kept bleeding.

That asymmetry is the Barro finding visible in a single table, and it is why “everyone declined” is no longer an adequate defense.

Here is where the traditionalist case, having gained real ground, runs into trouble. If the Missal is the mechanism, then wherever the Missal went, the collapse should follow. It did not.

Ireland received the Novus Ordo in 1970, exactly as America did. Irish weekly Mass attendance in the mid-1970s was around 91%. It was still around 87% in the mid-1980s. It did not fall below 80% until roughly 1990, did not reach 66% until 1997, and did not collapse to its present ~27% until the 2000s and 2010s.

That is a twenty-year lag between the alleged cause and the alleged effect, and the collapse, when it came, tracked the Celtic Tiger and the clerical abuse revelations with uncomfortable precision. A cause that waits two decades and then fires exactly when two unrelated shocks arrive is not behaving like a cause.

The Dutch case is the mirror image, and it is just as awkward for the liturgical thesis, because the Dutch collapse ran ahead of the Missal, not behind it. Dutch Catholicism in the 1960s underwent its own doctrinal upheaval: the Dutch Catechism of 1966, the national “Pastoral Council” of 1966–70 with its votes on clerical celibacy and contraception, and an open confrontation with Rome over the content of the faith itself. Attendance halved between 1965 and 1975.

Ireland and the Netherlands received the identical rite within months of each other and produced opposite curves. Whatever explains that difference, it is not in the Missal. It is in what each national Church was doing with its doctrine and its authority.

African Catholics numbered about 45 million in 1970. By 2024 there were more than 288 million. Africa’s share of world Catholicism went from 6.8% to 20.3% while Europe’s fell from 38.5% to about 20.1%, in the same decades, under the same Missal, promulgated by the same authority.

The reply that African growth is “just demography” is half right and worth stating precisely, because it cuts both ways. Africa’s total population roughly quadrupled over this period while its Catholic population grew more than sixfold, so there are real evangelistic gains beyond the birth rate, but the dominant term is indeed demographic. Which is exactly the point: if demography and social conditions can swamp the liturgy in Lagos, they can swamp it in Rotterdam. You cannot invoke confounders to explain away the growth and then ignore them when explaining the decline.

One further observation, since it is usually missed by both sides: African Catholicism is not liturgically minimalist. It is long, ceremonially rich, reverent, and strongly supernatural in its assumptions. The African case does not vindicate a stripped-down liturgy. It vindicates a serious one, which is a point the traditionalist should be making rather than avoiding.

The mainline Protestant comparison no longer proves what it was once thought to prove. Barro and colleagues have shown the Catholic decline is steeper and more persistent than a shared-culture story predicts. But the magnitudes still deserve to be seen, because they establish the size of the general force that any Catholic-specific explanation has to be layered on top of.

The growth is real and the demographics are genuinely striking, but the scale claim needs discipline, and the two available estimates disagree by an order of magnitude.

Pew’s 2023–24 Religious Landscape Study found 2% of US Catholics saying they attend a Traditional Latin Mass weekly, which would imply somewhere around a million people. The supply-side count says otherwise: roughly 500 non-SSPX parishes offering the older rite after Traditionis Custodes, plus about 103 SSPX chapels, at observed average attendances of roughly 150–200. That arithmetic yields something closer to 100,000–150,000 weekly.

For scale: CARA’s 2023 figures put weekly attendance at 18.6% of roughly 75 million self-identified Catholics, or about 14 million people at Mass on a given Sunday. The older rite is therefore drawing on the order of one percent of American Mass-goers, a real and growing constituency, and a very small one.

The gap is almost certainly survey error: respondents counting a reverent, ad orientem, or partly-Latin Novus Ordo as “the Latin Mass.” The supply-side number is the more trustworthy one, and honesty requires using it.

What the data does support is more interesting than raw scale. Bullivant and Cranney’s survey of TLM attendees found 98% affirming the Real Presence, against 31% of Catholics generally in Pew’s 2019 measure, and even against the more favorable CARA instrument, which found 64% using better-designed questions, the gap is enormous. TLM communities skew young, fertile, and catechized.

But this is a selection effect, and it cannot bear the weight put on it. People who drive forty minutes past three parishes to reach a Latin Mass are pre-selected for exactly the commitment being measured. You cannot infer from their zeal what would happen if the same rite were imposed on the median suburban parish, which is precisely the inference the argument requires.

Everything above concerns attendance, and attendance turns out to be the weakest ground the traditionalist can choose. There is a stronger argument, and it does not need a single statistic. Our companion piece Organic Development vs. Fabricated Liturgy sets out the Ratzinger critique at length; what follows puts numbers behind it.

It can be put in one sentence: the old rite has no author.

Nobody wrote it. There is no committee, no date of composition, no draft, no vote. The Roman Canon is substantially attested in the De Sacramentis of St. Ambrose around 390, and it stood essentially unchanged in its structure from the time of Gregory the Great, who died in 604, until the twentieth century. Around that core the rite accreted (prayers, gestures, feasts, sequences) the way a language accretes, by use rather than by design. When St. Pius V issued Quo Primum in 1570, he was not composing a Mass. He was codifying and pruning a Roman use that already existed, which is why the same bull left untouched every rite that could show two hundred years of standing. A legislator creating something new does not write in an exemption for older custom. A legislator recognizing custom does.

Set against that:

The Roman Rite. class=”cmp-body”>No author. No date of composition. Core prayer attested c. 390, structurally fixed by c. 604. Codified, not created, in 1570. Roughly fourteen centuries of accretion.

The Missal of 1970. class=”cmp-body”>Consilium established 25 January 1964. Forty-two members, later fifty-one, with over two hundred consultors, under Fr. Annibale Bugnini as secretary. Schemata, drafts, votes, minutes. Promulgated 1969. Roughly five years.

That contrast is not a matter of taste, and it does not depend on anyone’s attendance figures. It is a plain fact about how the two books came to exist. No rite in the history of the Latin Church had previously been produced that way, assembled to a brief, on a schedule, by a body convened for the purpose.

And the scale of the replacement is now a matter of counted record rather than impression. Matthew Hazell worked through the orations (the collects, secrets, and postcommunions) of both books line by line. Of the 1,269 unique orations in the 1962 Missal, 656 do not appear in the reformed Missal in any form at all. Of the 613 that were carried over, only about 17% came through with their text unchanged.

Roughly eight percent of the old Missal’s prayers passed into the new one unaltered. That number deserves emphasis for a reason beyond itself: Hazell compiled it specifically to test the claim, made by Archbishop Arthur Roche and earlier by Bugnini himself, that “almost all”, or on one telling ninety percent, of the old Missal’s prayers were retained. The manuscripts do not support it. When the Church’s own liturgical authorities describe the reform’s continuity with what came before, they have been describing something the texts do not show.

Lauren Pristas’s study of the collects reaches the same place from the other direction, by reading rather than counting. Every Sunday collect of Lent except Palm Sunday was replaced. Her conclusion about the Advent prayers is that where the older orations conveyed the absolute necessity of grace in the compressed manner proper to orations, the new ones “neither articulate it nor, more worrisomely, seem to assume it.”

Cardinal Ratzinger named exactly this, and named it as the problem:

In the place of liturgy as the fruit of development came fabricated liturgy. We abandoned the organic, living process of growth and development over centuries, and replaced it — as in a manufacturing process — with a fabrication, a banal on-the-spot product. Joseph Ratzinger, preface to the French edition of Klaus Gamber, The Reform of the Roman Liturgy, 1992

What makes this more than one cardinal’s opinion is that the Church had already condemned the underlying method, in a magisterial document, seventeen years before the Consilium first sat. Pius XII’s Mediator Dei (1947) rejected what it called antiquarianism: the assumption that older is automatically better, and that one may leap back over centuries of development to recover a supposed pristine form. It specifically refused the restoration of the altar “to its ancient form of table,” and the removal of the suffering Christ from crucifixes. That is a direct magisterial judgment on the archaeological method, and much of the 1970 Missal was built by precisely that method.

The Council itself said the same thing in Sacrosanctum Concilium 23: there must be no innovations unless the good of the Church genuinely requires them, and new forms must “grow organically from forms already existing.” Whatever else the Consilium’s five-year project was, it was not organic growth from existing forms. On this point the traditionalist is not arguing against the Council. He is holding the reform to the Council’s own stated standard, and the reform does not meet it.

So the charge stands, in its precise form. Not that the new Mass is invalid — it is not. Not that it emptied the pews, the evidence will not carry that. But that a rite which had never been authored was replaced by one that was, inside five years, by a method the Magisterium had already warned against and against a conciliar instruction requiring organic growth. That is a rupture in the Church’s manner of holding her own tradition, and there is no getting around it by pointing at Africa or at attendance statistics, because it was never a claim about attendance in the first place.

Put it all together and a coherent picture emerges; one that gives each side something real and takes something away from each.

The Council marks a genuine, Catholic-specific rupture. This is now the best-supported empirical claim in the debate, and it belongs to the traditionalist. The relative decline is measurable, it is significant, it has no pre-trend, and it survives comparison to Protestant and Orthodox countries. “Everyone declined, it was the sixties” is no longer a sufficient answer, and defenders of the reform who are still making it are behind the literature.

But the Missal is not the identified mechanism, and three natural experiments cut against it. Ireland kept full churches for twenty years under the new rite. The Netherlands emptied under its own doctrinal crisis. Africa grew sixfold. The same books produced opposite outcomes depending on what else was happening, which is the signature of a variable that is not doing the causal work.

What the evidence points to instead is what the study’s own authors point to, and what Greeley argued from the American data half a century ago: the shock was to authority and immutability. A Church that had grounded its claim on not changing, changed, visibly, rapidly, and on matters people had been told were settled. Humanae Vitae and the public dissent that met it did more measurable damage to Sunday attendance than anything in the Missal.

On that reading the liturgical reform mattered enormously, but as the most legible weekly proof that the immovable had moved, not as the mechanism of departure. It was the sign, not the cause. That is why the sisters vanished immediately and the priests took twenty years, why Ireland lagged and the Netherlands ran ahead, and why the same rite could preside over collapse in Europe and expansion in Africa.

This costs the traditionalist a rhetorical weapon he should not have been using, and it costs the reform’s defender the shelter of “you cannot prove causation”, which was never a defense of the reform’s fidelity to begin with.

And it relocates the argument onto the ground set out above, where it is far stronger. A rite that had never been authored was replaced, in five years, by one that was. Sacrosanctum Concilium required Latin to be retained and chant to hold pride of place, and both effectively vanished. The Council required new forms to grow organically from existing ones, and the Consilium’s project did not. More than half the old Missal’s orations were dropped outright, and of those retained only about one in six survived unedited.

Those are claims about fidelity, and they are true or false regardless of what happened to attendance. They were always the better argument. The statistics were never needed to make them — and, as it turns out, the statistics were never quite able to.

This article is the statistical half of an argument the site treats at length elsewhere. If you came for the numbers, these are where the theology and the history live.