Pretty good. Like Bok's "Lying", it assembles a good collection of data and problems about truth-telling and lying. Like Bok, he's in favour of taking honesty more seriously.
Here's a thing, though. He spends a lot of time gently berating those who are more concerned with what feels right than with what is factually true (prominent examples including Presidents Clinton and Bush-II). And then he does exactly the same himself! His closing chapter begins: "At the outset... we wondered whether more lies than ever are being told. In a sense, the answer to that question is beside the point. Because if we feel more lies are being told - and we obviously do - the effect is the same regardless of whether that feeling is valid" - in other words, regardless of the factual truth involved.
This seems a bit paradoxical, or self-undermining. But also the specific claim he makes here seems wrong, to me. If it were to turn out that in fact there were proportionally less lying going on nowadays, and that nevertheless we felt the opposite to be the case, that would be a new fact (or two) crying out for explanation, not one to be ignored in favour of just "believing our feelings".
That said, this also begs two perennially unanswered questions: (1) why do all these writers on the alleged decline of truth (Keyes, Bok, Oborne, Frankfurt, and the rest) feel so sure that there has been a decline in respect for truth in modern times? (Many of these writers even give potted histories of attitudes to lying, which show that toleration of some lying goes back well into the pre-Christian era). And (2) a methodological question: how on earth would such a curve be measured? How do you count lies, and how do you map the count from the days of majority non-literacy to the present, and from the small and geographically fragmentary world of recorded history (bit of Greece, bit of Rome, bit of Egypt, bit of Norse legend) to the hugely documented modern era?
I hate to be defeatist but it seems to me a task that one could only succeed in via a bunch of convenient idealisations (that is, white lies).
If so, it might be helpful for the truth'n'lies brigade to abandon the historical questions as both unanswerable and, for many purposes, irrelevant, and focus instead on the stakes involved in truth-telling and lying (and the other relevant phenomena) from a general and not a historical point of view.