The conventional wisdom is that this is a function of ideology. Conservatives say that the public is disgusted with liberal theology and marshal convincing statistics to support that claim. A variant on that description is the candle burning at both ends theory, that says that mainline churches are losing members both to conservatives on the right, and to secularism on the left.
But, a recent study from an Episcopal demographer begs to differ with this assessment. In this view, it mostly comes down to differences in birth rates.
The membership of the Episcopal Church has been predominantly white. And the birthrate among whites has declined substantially since the fifties. Hathaway finds (p. 13) “the association between [white birthrate and mainline membership growth] is so strong that it produces a correlation of .94 (0 being no relationship and 1.0 begin a perfect relationship). In statistical terms, 88% of the year to year variation in mainline membership can be explained by the birth rate.” (He finds the correlation .89 between white birthrate growth and member growth in the Episcopal Church.) Further (p. 16),
"As noted earlier, all denominations—mainline and conservative—were affected adversely by social changes occurring in the 1960s and 1970s. However, mainline denominations were hit hardest by the changes because declines in the birth rate were much more severe among the more highly educated white population. (Among conservative Protestants and Mormons the birth rate remained much higher than for the mainline, insulating these groups from the full effect of declines in fertility)."
and (p. 17)
"The Episcopal Church has the highest proportion of members among mainline denominations who are college graduates and in households earning $75,000 or more. As a result, the birth rate among Episcopalians is much lower than the national average—and even lower than the population of non-Hispanic whites. A reasonable estimate, based on education and race, is approximately 1.5 children per woman (compared to the replacement level of 2.1) for Episcopalians."
The evidence for the demographic view is convincing, and also explains why one would see in a single Roman Catholic Church with a single ideology, a decline in Anglo membership matched by an increase in Hispanic membership, to produce little net change.
The Wall Street Journal's Op-Ed page noted data that support a mixed demographic and political view a year and a half ago, implying that the demographic differences between denominations that fall upon political lines are no coincidences:
According to the 2004 General Social Survey, if you picked 100 unrelated politically liberal adults at random, you would find that they had, between them, 147 children. If you picked 100 conservatives, you would find 208 kids. That's a "fertility gap" of 41%. Given that about 80% of people with an identifiable party preference grow up to vote the same way as their parents, this gap translates into lots more little Republicans than little Democrats to vote in future elections. Over the past 30 years this gap has not been below 20%[.] . . .
The fertility gap doesn't budge when we correct for factors like age, income, education, sex, race--or even religion. Indeed, if a conservative and a liberal are identical in all these ways, the liberal will still be 19 percentage points more likely to be childless than the conservative. Some believe the gap reflects an authentic cultural difference between left and right in America today. As one liberal columnist in a major paper graphically put it, "Maybe the scales are tipping to the neoconservative, homogenous right in our culture simply because they tend not to give much of a damn for the ramifications of wanton breeding and environmental destruction and pious sanctimony, whereas those on the left actually seem to give a whit for the health of the planet and the dire effects of overpopulation."
The Episcopal demographic study doesn't directly measure either inflows through conversion to mainline denominations, or outflows by people who leave religion entirely or join a different religious denomination. But, looking at the birth rate data and bottom line numbers, the implication of the demographic study is that conversions into and out of denominations are largely a wash.
A Baptist source (cited by an Episcopalian and probably reporting on a study also reported here) puts the relative contributions of birth rate and conversion at 70% birth rate and 30% conversion, while a labor source cited in the same place acknowledges but down plays the denominational impact on birth rates.
Christian Century magazine has a 2004 report that is more nuanced.
Membership statistics have nuances that support a variety of conclusions. If Unitarians, a liberal denomination, grew by 80 percent between 1963 and 1976, if liberal Protestants and Jews have low birth rates while conservative Protestants have high birth rates, if church attendance across denominations has been relatively constant in the 20th century, then the received notion of liberal or mainline "decline" since the 1960s needs some complicating footnotes. In any event, denominational religion's twists and turns provide grist for the historian's mill, enhancing an overall understanding of the evolution of the American religious scene that includes a shift toward nondenominational religious expression.
One place where this doesn't add up is the growing ranks of the not religious. Non-religious people are now a much larger share of the population than they used to be, but are demographically closer to the mainline Christian denominations than to the conservative Protestant and significantly non-Anglo denominations. This can't be explained with birth rates.