The estimate was he was doing more more than 20MPH which resulted in them rapidly approaching and overtaking the vehicle they suddenly encountered which they later realized was Jeremy.
That's incorrect. Their original estimates were "about 30 mph", "about 30 mph" and "no more than 30 mph". They never estimated "more more than 20MPH". They were travelling at high speed, so their approach was rapid even if Jeremy was travelling at 30 mph. Even at very high speed, they cannot overtake instantaneously - it takes a second or two (which is "a period of time", albeit a short one). The only way of calculating 20 mph as a maximum speed for the last mile is by assuming it took Jeremy at least 3 minutes, but the "3 to 4 minutes" that the police put in their statements was not based on use of a watch - it was just an estimate they agreed on that (in hindsight) conflicted with their estimates of Jeremy's speed. There was no reason for deciding that their "3 to 4 minutes" estimate was more accurate than their "about 30 mph" estimate.
. . . in fact both of you present the same made up lie that Jeremy phoned Witham, then Julie and then Chelmsford
That's mike tesko's theory, but we can't be sure that Jeremy tried to call Witham, as that may have been something that the police invented later as having been mentioned by Jeremy.
They put NO MORE than 30MPH in their statements
That's incorrect - only Pc Saxby used that wording; the other two officers used "about 30 mph".
. . . they drove the distance and took into account the distance they recalled between their speed and his speed.
That doesn't even make sense. The police clearly didn't try to calculate a relative speed (their own speed was irrelevant); they just realized that taking at least 3 minutes to cover a mile implies an average speed of at most 20 mph, but "at least 3 minutes" wasn't known for certain, so their conclusion was unsound. All they knew was that their time estimate and their speed estimate weren't both accurate. In my opinion, it's much easier to estimate reasonably accurately the speed of a car as you overtake it than to estimate accurately without use of a watch or clock the length of a time interval of several minutes, especially when there's no particular reason for estimating it.