I agree with meta... the numbers can't be trusted. They just don't make sense. Focusing just on the U.S. numbers, I look around at several congos I am familiar with and over the past 10 years there has been a serious decline in their membership. The ancedotal evidence is plain. In the 10 years between 1985 and 1995, there was plenty of growth to be seen. Congos were dividing and growing, then dividing again. It was not uncommon to see KHs that held one congo in the late '70s now have three or four groups sharing the building.
Beginning in 1995 (a "marked year," in my opinion, thanks to the "adjustment" in dub understanding of the generation doctrine), that all stopped. In our area, there were three congos in our local KH in that year. There are still only three, 10 years later! During one period of 18 months, just before we finally left, one of the congos went from 133 cards in the file (publishers) to 84, while the other two remained about static (so those who went missing didn't just move over to another congo). And this is in an area whose population is one of the fastest growing in the country.
To offset what must be a troubling set of statistics, there is no doubt peer pressure has led local elders to fudge the numbers anyway they can. I don't mean they make up the numbers, but there are ways. Call all the inactive ones and coax them into reporting one hour, for example. Look the other way when people turn in reports that are clearly bogus. It's all about retaining "publishers," even if in name only. Some posted this:
Even 15 minutes per month will now put you in the count. So, if they beat the bushes harder, a few more publishers straggle in...
That's been a real bonus for the WTS. In the last elder's school I attended, the CO introduced this new "program" and said the Society had piloted it in a couple of foreign countries and the result was a 2% increase in publishers. Looking over the numbers in the past couple of years, that 2% alone is the difference between a national report with a positive bottom line vs. one in negative territory.