Preface: There are a lot of assumptions about numbers in this post. I have explained the assumptions made and why, but of course you could massage the numbers to paint a much different picture if you wanted to. I have tried to err on the side of overestimating the situation rather than catastrophizing it (eg. by positing that it is still possible to stabilize at this point and by not even attempting to account for general population inbreeding). But even with these generous assumptions, Hokkaido are in a lot of trouble.
How bad is the situation for Hokkaido? How many Hokkaido are there in the world? Those are difficult questions to answer. As of October 2020, there are 55 Hokkaido in North America and probably a bit more than that in Europe. But the majority of the population lives in Japan. Since none of the populations outside of Japan are large or diverse enough to be sustainable without Japanese imports, and every export to a foreign country reduces the number of dogs in Japan, let's just consider the Japanese population to be the world population.
Okay, so how many Hokkaido are there in Japan? Well, again, we don't have an exact answer. There are no data sources on the number of living Hokkaido, nor any official records of Hokkaido deaths to go by. But we do know how many are born and registered each year in Japan and approximately how long Hokkaido live. As long as the birth rate remains constant and no new natural disasters befall Japan (such as the Great East Japan Earthquake in 2011) the death rate will remain fairly constant as well.
But the birth rate has been falling. For the last few years Hokkaido registrations have been under 300 per year. In 2018, there were 242 Hokkaido registered. That is a 64% decline from the 679 registered in 2008, and an 83% decline from the 1,389 registered in 1993. If this trend continues, Hokkaido will be gone completely within the next few years or decades.
But let's assume we manage to stabilize the breed at a constant 250 born per year and use that to estimate the population size. How many Hokkaido would that give us? Well, Hokkaido live about 15 years on average. So looking ahead to 15 years from now, when most of the dogs alive today will have been replaced by future generations, that would give us 15 × 250 = 3,750 living Hokkaido. Again, this will remain constant as long as the birth and death rates are constant, since new pups will be born to replace those that die.
Seems like we could manage with that, right? However, it is important to remember that those are not 3,750 diverse, unrelated individuals. With an average litter size of 4 puppies, the 250 Hokkaido born per year are actually the result of 60-65 litters. The majority of pups born each year have siblings who are extremely genetically similar to themselves. That means that, effectively, the number of new genetic combinations — even without considering how inbred the parents are or whether it was a repeat breeding — is only 60-65 per year and that the effective population of individuals with distinct genetic combinations is only 25% of the living population.
Also keep in mind that those are not 3,750 breeding dogs. Breeding age is typically from 2 to 7 years old, which means that roughly 60% of the Hokkaido population will be unbreedable in any given year due to being too young or too old. Of the remaining 40% who are of breeding age, the vast majority will not be bred that year. If we generously assume 10% of all breeding-age Hokkaido are bred each year, that means our active breeding population for any year is only 4% of the living population. (In truth, the number of breeding-age dogs who are bred each year is probably much less than 10%.)
This means that the 250 dogs per year are coming from only 4% of the population, and 4% × 3,750 = 150 breeding Hokkaido. If you adjust for full siblings having effectively the same genetics, those 150 dogs are really the equivalent to just 38 distinct individuals. 7 years from now, all of the breeding Hokkaido will be dogs born in the previous 2-7 years from those 38 "individuals". Another 7 years will see it reduced by the same amount again. In other words, the gene pool narrows considerably every year even with a constant birth rate keeping the total population numbers up.
Stopping the decline in births is therefor not enough. The only way to save the breed at this point is to dramatically increase the births and increase the percentage of breeding-age dogs who are bred each year. We do also need to increase the number of new genetic combinations by eschewing repeat breedings, make sure at least one pup from each combination goes on to be bred, and de-prioritize breeding the full siblings of dogs who have already been bred repeatedly. But before we can worry about any of those secondary (though very important) concerns for diversity, we must get the birth rate up. We're losing a ton of genetics every single year that it remains constant or declines.
Afterword on the Term "Endangered"
I will often refer to the Hokkaido as an endangered breed. By population size it may seem that they're not endangered — after all, I am predicting that in 15 years there will be still be 3-4k Hokkaido in the world. But it is critical to remember that humans control the breeding of domestic dogs, who by and large reproduce far less often than wild animals do. In the wild, animals will breed every breeding season they're capable of it, and most individuals who do not have debilitating issues will reproduce. By contrast, the vast majority of dogs are never bred at all, and even those who do reproduce are not bred as often or as long as they're able. Dogs breed and exist solely at the whim of humans, and larger population numbers compared to wild animals can therefor be deceptive.