This is indeed a very good question, that I have spent a long time thinking about myself. My take on this is that there is indeed a very close relationship between inbreeding and speciation, but that inbreeding actually PRECEDES speciation ! The key to this rather counter-intuitive point of view is that inbreeding actually has several advantages, including that of resulting in cleaner genomes in those offspring that survive inbreeding depression.
In populations that have high degrees of inbreeding, because of, for example, small sizes of populations, or high tendencies to self fertilisation, the recessive mutations that cause inbreeding depression will be progressively eliminated, and there will consequently be very little inbreeding depression.
In such a context, there will thus be no barrier to exploiting other advantages of inbreeding such as reducing the cost of sex, or keeping together advantageous gene combinations. Small groups of individuals would thus be better off breeding among one another than with the ancestral stock, leading to speciation.
If you are interested, or just intrigued, by this kind of concept, I invite you to read the rather long essay I wrote on the subject (The existence of species rests on a metastable equilibrium between inbreeding and outbreeding. An essay on the close relationship between speciation, inbreeding and recessive mutations, Etienne Joly, http://www.biologydirect.com/content/6/1/62 ).
And please do not hesitate to get in touch with me directly by email ( atnjoly(at)mac.com ) if you have comments or any further questions after reading this essay.