Roderick Floud puts the case for the Retention of Personal Records
Historians are ambivalent about secrets. Nothing is more exciting than discovering one, digging up a hoard which was buried to escape an invader or finding a crucial letter tucked into the binding of a book. So it is good that secrets exist. On the other hand, if a secret is too well hidden, it is unlikely that it will ever be revealed.
The ideal secret, then, is one that is well hidden but not too difficult to find when the historian later looks hard for it. Sometimes this happens by chance: I remember discovering with delight an exchange of letters about price-fixing on government contracts. The letters had survived only because the firms involved had wanted to keep them secret; so they were kept separately from the other records which had been destroyed.