Genetic diversity cannot decline when you start from a single pair. With sexual species, a single pair is as low in diversity as you can get without going extinct. Diversity can only increase through mutation, so since mutational effects are generally minimal over short time periods one would not think that diversity could possibly increase.
This quite wrong. In fact, a decline is exactly what you'd expect.
To see why, let's consider a single allele, and suppose there's initially maximal genetic variation at this allele - that is, both founders are heterozygotes and don't share any of the variants. Now, either both can pass both of their gene variants onto a least one of their offspring, or one of the variants can be lost. Now, presuming there's a reasonable number of offspring the chance of any particular gene variant being lost is quite low, but there's going to be a few thousands such alleles which meant that even the relatively unlikely will happen a few hundred times.
In the next generation, things get worse, because now the only available mates are brothers and sisters and that means that the chance of the offspring being homozygotes sky-rockets.