Almost all types of diabetes have a known genetic component. A very small number of diabetes types are determined by one gene- you got the gene, you got the diabetes. A much larger portion of diabetes types and cases are multifactorial in cause, but also in the specific genetic cause.
Meaning, as a type 1A diabetic, that my likelihood of developing diabetes was strongly impacted by MULTIPLE genes in the region of the genetic code where there is more variability in genetics, but that some factors other than my genetics certainly impacted the fact that I got diabetes- if type 1A diabetes was totally genetic, identical twins would always both have it or both not, and that's not how it goes.
Anyways, one of the interesting things about the genetics of type 1 diabetes is that the highest risk genes are heterozygous. That means that, unlike in straight up genetic diseases where having two parents carrying the same faulty allele (carriers) resulting in an increased chance of the disease, in type 1 diabetes, the highest risk comes for people who have combinations of different high risk alleles, not two of the same high risk allele (an allele makes up one parent's contribution towards a gene). If this paragraph didn't make sense to you, don't worry about it. Do ask me to clarify if you know how to get me to word it better though.
Anyways. One of the more unusual explanations for increases and decreases in disease are based on which genes are becoming more and less common. This is based partly on who has how many kids (that determines allele frequency), but also on who people have their kids with. It should be obvious that certain diseases (like the above mentioned recessive genetic disorders) are less common if people are having their kids with people who are not their fourth cousins, etc. With more complex genetically influenced diseases (including most autoimmune diseases), it's a little harder to predict what population mobility will do to disease patterns.
However, one of the things that seems to be happening is that, more so than in previous generations, people are choosing to have children with partners whose alleles in the areas that code for the immune system are complementary- they are picking to have kids with people whose immune systems are different from their own. This is mostly a good thing (at least according to theory) but it is resulting in more people of my generation and below who have the high risk genes for developing type 1 diabetes.
Unfortunately, this theory is not supported by the American evidence, which says that among those diagnosed with diabetes as kids in the last many years, fewer these days have the high risk immune system genes. I was looking for the article on which I based most of this post, and found instead some studies comparing the frequency of the highest type 1 diabetes risk genes in the US and saying that it's gone down, and now I sort of feel like erasing what I wrote, but I won't. Instead I'll give you a link to this excellent article and count the fact of the day to be as follows:
The portion of children diagnosed with diabetes in the United States whose diabetes can be attributed to high genetic risk declined between 1965 and 2005.