A few steps ahead of EPO
Synthetic EPO in theory became available for athletic use in 1987, when the world record in the men's 1,500m stood at 3:29.46 (1985, Said Aouita, Morocco). Hicham El Guerrouj's current mark, which has stood since 1998, a drop of 1.65 percent from the pre-EPO era. In reality, no one knows exactly when EPO became a major thing in distance running, but you can be certain that athlete managers were hunting for it the moment they learned it could be made in labs as well as in kidneys.
Aouita also held the outdoor 3,000m record for a spell, being the first to dip under seven and a half minutes (7:29.45, 1989), breaking Henry Rono's 11-year-old record by over two and a half seconds. If I had to guess, which I obviously do, I would say that Aouita was probably the last world record holder in a distance event who can be almost definitely removed from EPO suspicion on logistical grounds alone, which isn't to say I think he was any cleaner by the standards of his day than anyone else. In any case, Aouita's 3,000m record has dropped by 1.95 percent. The record (7:20.67) has also been static since Daniel Komen set it in 1996, and in fact hasn't been seriously threatened. (I think Yomif Kejelcha has as good a shot as anyone has in the past 20 years now that Kenenisa Bekele has missed his chance.)
The 3,000m steeplechase record has fallen from 8:05.35 in 1989 to 7:53.63 today, with hat mark now fifteen years old, although it seems unfair to discount Brahim Boulami's 7:53.18 from 2002, since it's known he was on EPO. That's a drop of 2.39 percent.
The 5,000m record fell from Aouita's 12:58.39 in 1987 to Bekele's 12:37.35 in 2005, s drop of 2.70 percent, and the 10,000m mark was trimmed from Arturo Barrios' 27:08.23 in 1989 to Bekele's 26:17.53 in 2005, an improvement of 3.11 percent. Both records still stand.
The marathon is an outlier here, and not merely for being a road race and involving unique physiological demands compared to the aforementioned events. There is also a great deal more financial incentive at the world-class level, in large part because of the introduction of the World Marathon Majors in 2006.
Surprisingly, the world record in the marathon did not change during the 1990s until Ronaldo da Costa ran 2:06:05 in 1998 to take 45 seconds off the record set by the insanely anonymous Belayneh Dinsamo. The longest period without a new record since 1998 is 4 years and 2 days. Eliud Kipchoge's official mark of 2:01:39 from 2018 -- which Bekele came within two seconds of matching last month on the same course in Berlin -- is 4.09 percent faster than Dinsamo's.
I am hoping at this point that putting these distances in ascending order makes it clear that the records in the men's distance events have fallen by greater amounts with increasing race distance. There are a number of obvious issues with this rough assessment, among them the fact that the records listed aren't all from the same time frame (outliers in any sample will do that); otherwise, a graph no one will look at would be useful here. If one attempts to account for this to some extent by using Tergat's marathon record from 2003 (2:04:55), the drop from Dinsamo's record is only 1.51 percent, and the improvement level seen between the late 1980s and the mid-"oughts" in the 10,000 (about 3 percent) wasn't observed in the marathon until 2014, when Dennis Kimetto became the first man under 2:03:00.
I think the running world, of which I remain a mostly cognizant part, is coming around to the fact that the latest racing shoes really can make a phenomenal difference in the marathon on the right set of feet. That last disclaimer does a great deal of work, because most people who run marathons would be extremely ill-advised to run marathons in ordinary flats, much less something with zero cushioning at all. The runners who benefit most from the VaporFly 4% shoe are most likely those who are already extremely efficient, making this a case of the rich getting richer.
I don't know how much a factor EPO is on the roads, but I can say with confidence that most of the improvement in elite marathon times (and the improvement of the all-time top lists down to whatever ranking you pick) in recent years have been owed more to the shoes then to the drugs. I don't know if running will ever confront the issue of whether such footwear ought to be judged an unfair advantage, but unless they actually provide electrical power or something, they represent nothing more than one more clever engineering innovation. A few people my age will eventually start blare on about how much faster they would have been if better shoes had been available in their day, and anyone who hangs with other runners probably already knows which goon or two in the group will be the first to do so.