A couple of days ago the Office of the General Assembly (OGA) of the Presbyterian Church (U.S.A.) released the Comparative Summaries of Statistics for 2019 and the corresponding press release and narrative. This does, of course, provide new data points for my growing data set and gives me an opportunity for some statistical analysis which is, for me, a “source of innocent merriment.”
In a general sense, the results are predictable – The membership of the two branches that would become the PC(USA) in 1983 peaked in 1965 at 4.25 million members and has been declining ever since. In terms of the direction of change in 2019, it is still a decline. But the good news is that the decline rate has been decreasing. What that means we will get to in a minute.
Specifically, the PC(USA) finished 2019 with 1,302,043 members, a net decrease of 50,635 or 3.7%. There were 45,079 members added to the rolls, roughly half through Professions of Faith and Reaffirmations. There were 91,552 members lost, with about 13% transferring to other churches, 26% moved to the Church Triumphant, and the bulk, almost two-thirds or 56,133 members were deleted from the rolls for other reasons.
The number of churches decreased by 120, or 1.3%, to 9041. The statistics show that 19 were organized, 95 were closed, none transferred in, and 24 transferred out to other denominations. There are also 110 New Church Developments/Fellowships – a decrease of six, and 176 New Worshipping Communities – an increase of 18.
So what does this look like in the context of the last 27 years:
| Year | Num Churches | Num Change | % Change | Num Members | Num Change | % Change | Mem/Church |
| 1993 | 11,416 | -40 | -0.3% | 2,742,192 | -38,214 | -1.4% | 240.2 |
| 1994 | 11,399 | -17 | -0.1% | 2,698,262 | -43,930 | -1.6% | 236.7 |
| 1995 | 11,361 | -38 | -0.3% | 2,665,276 | -32,986 | -1.2% | 234.6 |
| 1996 | 11,328 | -33 | -0.3% | 2,631,466 | -33,810 | -1.3% | 232.3 |
| 1997 | 11,295 | -33 | -0.3% | 2,609,191 | -22,275 | -0.8% | 231.0 |
| 1998 | 11,260 | -35 | -0.3% | 2,587,674 | -21,517 | -0.8% | 229.8 |
| 1999 | 11,216 | -44 | -0.4% | 2,560,201 | -27,473 | -1.1% | 228.3 |
| 2000 | 11,178 | -38 | -0.3% | 2,525,330 | -34,871 | -1.4% | 225.9 |
| 2001 | 11,141 | -37 | -0.3% | 2,493,781 | -31,549 | -1.2% | 223.8 |
| 2002 | 11,097 | -44 | -0.4% | 2,451,969 | -41,812 | -1.7% | 221.0 |
| 2003 | 11,064 | -33 | -0.3% | 2,405,311 | -46,658 | -1.9% | 217.4 |
| 2004 | 11,019 | -45 | -0.4% | 2,362,136 | -43,175 | -1.8% | 214.4 |
| 2005 | 10,959 | -60 | -0.5% | 2,313,662 | -48,474 | -2.1% | 211.1 |
| 2006 | 10,903 | -56 | -0.5% | 2,267,118 | -46,544 | -2.0% | 207.9 |
| 2007 | 10,820 | -83 | -0.8% | 2,209,546 | -57,572 | -2.5% | 204.2 |
| 2008 | 10,751 | -69 | -0.6% | 2,140,165 | -69,381 | -3.1% | 199.1 |
| 2009 | 10,657 | -94 | -0.9% | 2,077,138 | -63,027 | -2.9% | 194.9 |
| 2010 | 10,560 | -97 | -0.9% | 2,016,091 | -61,047 | -2.9% | 190.9 |
| 2011 | 10,466 | -94 | -0.9% | 1,952,287 | -63,804 | -3.2% | 186.5 |
| 2012 | 10,262 | -204 | -1.9% | 1,849,496 | -102,791 | -5.3% | 180.2 |
| 2013 | 10,038 | -224 | -2.2% | 1,760,200 | -89,296 | -4.8% | 175.4 |
| 2014 | 9,829 | -209 | -2.1% | 1,667,767 | -92,433 | -5.2% | 169.7 |
| 2015 | 9,642 | -187 | -1.9% | 1,572,660 | -95,107 | -5.7% | 163.1 |
| 2016 | 9,451 | -191 | -2.0% | 1,482,767 | -89,893 | -5.7% | 156.9 |
| 2017 | 9,304 | -147 | -1.6% | 1,415,053 | -67,714 | -4.6% | 152.1 |
| 2018 | 9,161 | -143 | -1.5% | 1,352,678 | -62,375 | -4.4% | 147.7 |
| 2019 | 9,041 | -120 | -1.3% | 1,302,043 | -50,635 | -3.7% | 144.0 |
So what is one of the reactions to the 2019 data? The headline on the PC(USA) news article from the Office of the General Assembly is “PC(USA) statistics show a leveling off in membership decline.”
So yes, from the peaks in the 2016 numbers the rate of decline for both churches and members has shown a consistent decrease in the rate of decline. So things are getting better. But are they levelling off? Let’s look a little bit closer.
Both the number of churches and the number of members had a somewhat consistent decline for the first part of this time period through about 2004. The membership decline was creeping up but still hung below 2%/year. The rate of decline in the number of congregations was similar hanging at or below 0.4%/year. Both then show a bit of acceleration up to 2011 with the rate of church decline rising to just below 1%/year and the membership decline rising to a bit over 3%/year. Then in 2012, there was a rapid increase to a plateau that continues in the 2016 data. The rate of decline of the number of congregations was right around 2.0%/year and the decline in total membership was generally above 5%/year. Since 2016 the numbers have been dropping but still hang at numbers above the earlier periods on the chart. Here are the graphs for the number of churches and the number of members.
So is it leveling off? Well, it could be, but a good likely scenario is that it returns to the lower rates of the 1990’s.
Another reaction comes from the Stated Clerk J. Herbert Nelson. The OGA article includes his full statement concerning the 2019 statistics in which he begins by saying “For the first time in more than thirty years, the PC(USA) is not reporting membership losses. Our membership remains at 1.3 million. This is good news!“
So yes, if you work with two significant figures and truncate after the two significant figures that is a correct statement. But, if you use standard scientific practice for significant figures we have to remember that the 2018 number to three significant figures is 1.35×106 (1.35 million members). The rule for rounding is that if the first insignificant figure is a five it gets rounded down if the next number is even and up if the next number is odd. In other words, it gets rounded so that the last significant figure is even. So, in this case, the PC(USA) membership in 2018 to two significant figures is 1.4 million. (Remember, I am a science professor in my other life.)
However, one of the rules of significant figures is that exact numbers, which a count of the membership of the church would be, has all numbers significant so there is a difference between 1,352,678 and 1,302,043.
OK, now my turn to take the Rorschach test.
Looking at the numbers I see two things. First, I see a decline rate that is returning to mainline/oldline baseline. Over a decade ago I looked at the seven denominations that are classified as the mainline, or sometimes oldline, churches. At the time some of them, like the PC(USA) and the Episcopal Church were heavily embroiled in controversy. Others, like the American Baptist Church and at the time the United Methodist Church, had little internal controversy. My conclusion was that there was a common baseline decline in the branches of a bit less than 1% per year. In addition, the more controversy a denomination was involved with the higher the decline rate. For the PC(USA) this peaked a few years ago with a total decline of 5.7%/year so an additional decline of about 5%/year beyond the baseline rate.
It is probably time to return to that analysis, but the first thing I see, as I mentioned above, is not so much a levelling off of the PC(USA) decline rate but return to the slow steady decline of around 0.5%-1.0% of the late ’90s in the table above.
The second thing I see is in that right-hand column of the table. As the PC(USA) decline continues the church loses members at a proportionally higher rate than churches are closed. Between 1993 and 2019 the average number of members per church has dropped by almost 100 members/church from 240.2 members/church to 144.0 members/church. Graphically, the change looks like this.
To look at this from another point of view, if the 2019 membership change is thought of as the dismissal of 120 churches and their members in 2019, the average number of members dismissed in those 120 churches would be 422 members.
On the one hand there is an important place for small churches. They are important to small rural communities and for niche ministries and settings. However, a denomination full of small churches would challenge the sustainability of the Presbyterian concept of the graded councils from the local to the national level.
The Church of Scotland in the radical restructuring plan that was initiated two years ago and adopted last year realized this. This plan restructures the Kirk broadly so that it is more flexible and sustainable. This restructuring includes consolidating churches and merging presbyteries. But an important difference is that the Church of Scotland has authority for these decisions set at the national level while American Presbyterianism locates ultimate authority at the presbytery level. That makes broad decisions like this more challenging.
There are probably two more topics to mention briefly. The first is the rough projection of where the PC(USA) is headed. If indeed the church is headed towards the mainline decline trend, if in five years it reaches a church decline trend of 0.3%/yr and a membership decline of 1.0% that would suggest that in 50 years the PC(USA) will have about 7,500 churches and 730,000 members and so would have just under 100 members/church.
The other thing to mention is that the PC(USA)’s new model for ministry is not in these numbers. As mentioned earlier it reported 110 new church developments and fellowships in 2019 and 176 New Worshiping Communities. That is a slight decrease in the first and a slight increase in the second. The New Worshiping Communities (NWC) are dynamic, non-traditional forms of community. Many of these are pretty small, some are short-lived, and generally, their creative and dynamic nature means that they don’t have members in the typical sense of Presbyterian churches. So their attendee numbers are not reflected in the annual statistics.
So there we have this year’s statistical report with three views of what is seen in the numbers. Lots of other numbers in that Comparative Summary of Statistics if you want to take the Rorschach test and see what you see in the data. Have fun.
Interesting article!
My comments:
1) you wrote: “So is it leveling off? Well, it could be, but a good likely scenario is that it returns to the lower rates of the 1990’s.” I’m not sure why you think the 90’s are the norm. As you noted, the rates of membership loss gradually crept up to 1.8% in 2004, and then increased to 3.2% in 2011, and peaked from 2012 to 2018, dropping to 3.7% in 2019. But this 3.7% decline in 2019 is pretty much back on track with the trend from 1983 to 2011. The societal forces causing the steady decline from 1983 to 2011 are still at work. Competing ways to spend a Sunday morning are becoming more attractive, such as youth sports leagues, while the attractiveness of a church service is declining as attendance steadily drops. Choirs are getting thinner, youth programs are getting reduced, specialized contemporary vs. traditional service options are getting eliminated, etc.
2) you ignore Covid. The 2020 decline will almost certainly be worse than 2019, especially for smaller congregations that weren’t able to adapt to an online format and don’t have resources in reserve.
3) There are 1,612 congregations reporting fewer than 25 members in 2019, and 800 of those report 16 or fewer members. Some have as few as 1 reported member. Many have no reported website, and no reported income, and no reported pastor. It would not be surprising to see a wave of closures from this category alone over the next several years that exceeds the 120 annual closures in 2019.
4) You wrote: “The rule for rounding is …”. But there is no single rule for rounding. You suggest “round to even”, which is suitable for some purposes like reducing bias. The more common rule for rounding, as taught in school, is “round to nearest”, which suits other purposes like minimizing rounding error.
5) The problem with the 50,000 lost members that somehow turned into a claim of no losses to report doesn’t have to do so much with which rounding technique is used, it has more to do with when the rounding is applied. If you are asked to add 1.4+1.4+1.4 rounded to an integer, the most accurate way is to add first, getting 4.2, and then round to 4. The least accurate way is to round each 1.4 to 1 and add the three 1s, getting 3. The reason is that rounding introduces rounding error, and rounding error accumulates, so you want to round once at the end of a calculation rather than multiple times during the calculation. In this case, the 2018 reported membership was 1,352,678 and 2019 was 1,302,043. If you subtract first rather than rounding first, you get a difference of 50,635, which can be rounded to the nearest 10,000 to get 50,000 rather than zero.
6) The press release you cited has a similar rounding issue. It says: “The newly released statistical report for 2019 shows a total membership of 1.3 million, compared to 1.4 million three years ago, but the annual decline in churches and active members is lower than in years past.” Again, this is rounding first, purporting to show a 3-year decline of 0.1 million. But 3 years ago, the 2016 reported membership was 1,482,767, and 2019 was 1,302,043, a difference of 180,724. Rounding that to the nearest 0.1 million gives 0.2 million.
7) You make no mention of the reliability of the reported membership numbers. But churches vary in their urgency in purging their attendance rolls, and the current figures are only an approximation. The number of members who should have been removed from the rolls but weren’t is probably greater than the number who actually were removed, as indicated by reported average attendance that is 43% of membership.
8) You wrote: “if in five years it reaches a church decline trend of 0.3%/yr and a membership decline of 1.0% that would suggest that in 50 years the PC(USA) will have about 7,500 churches and 730,000 members and so would have just under 100 members/church.” Projecting the 2019 membership decline of 3.74% out 20 years leaves 607,495 members.
The reported number of members in the smallest church in each Synod for 2019 are shown below.
I’m sure each one of these has an interesting story behind it.
Alaska-Northwest 6
Puerto Rico 8
Lakes And Prairies 3
Lincoln Trails 3
Living Waters 1
Mid-America 1
Mid-Atlantic 1
S California And Hawaii 12
South Atlantic 2
The Covenant 1
The Northeast 2
The Pacific 2
The Rocky Mountains 5
The Southwest 4
The Sun 1
The Trinity 3
I don’t think your 50-year projection considers age demographics and its affect on projected membership declines and church closures.
In 2019, 31% of all members were older than 70, and another 25% were over 55.
In the smaller churches, which presumably are more at risk of closure, these percentages are higher, with 45% of reported members in churches with 1-25 members being over age 70.