If you want to model the global population, you need to use the parameters that are appropriate for the global population. If you wish to narrow down to a specific set of a population, then a set of parameters that are specific to that set of circumstances will apply.
The model is a biological model. It does not reflect the affect of extreme government policies. And the parameters can change for changes in technology. If I were to develop a cheap and efficient water purifier that could be distributed to all families in Calcutta, the population there would increase in a manner that could not be predicted with the old set of parameters. The population would increase until another factor limits population growth (such as food supplies). However, even a drastic change in technology will not affect the parameters too much. Likewise, if China decides to kill all children after the first one, the model will not reflect this.
If Verhulst's growth model was valid, why didn't the United Nations take it into consideration in the 1970s, when it was jumping up and down telling us the sky was falling? The reason is that the formula didn't hold weight back then. I don't think it does now. What use is it if you have to continually insert qualifiers.
What you are basicly saying is that it's a fabulous model as long as we can insert the "unknown factor" whenever it doesn't reflect reality. Here's the unkown factor for the US. Here's the unknown factor for China. Here's the unkown factor for Europe. Here's the unknown factor for the Middle-East. And here's the multi-faceted unkown factor for the planet. I have a rather strong suspicion that this formula is always going to be recognized by some people as extremely accurate after the fact, but will never be the stuff of an accurate predictor.
In one of my replies I made the observation that this planet's populace is still increasing at the rate of 100 million per year. In eighteen years that will be 1.8 billion. Add 1.8 billion to 6.2 billion and you have 8 billion. That's reality. I don't think the Verhulst growth model has exhibited it's reliability.