You're missing the point, he's saying the Japanese Government could shut down Gox in a permanent way and have their assets frozen for years. I have been wondering this same thing myself...
With all the exchanges being shut down around the world, why has the largest one never been targeted by anything more than unruly hackers? Gox is the single largest chokepoint and is entirely responsible for determining the price of BTC as most people see it, yet it seems immune to regulatory troubles that increasingly plague exchanges around the world. Is Japan such an advantageous legal environment to operate in? Are there other international exchanges taking advantage of this and operating out of Japan?
If Gox was shut down by the Japanese government it would have a MAJOR impact on the price, and while it might recover depending on how other exchanges react and how usable decentralized exchanges are by that point, I think it's obvious the impact would be probably as significant as anything else in Bitcoins lifecycle so far.
It's not about losing BTC to a hack, it's about a sea-change event eliminating the largest player in the currency conversion game
I have to think that most people do try to limit the BTC and fiat they have on account at Mt. Gox out of general principle. I did when I was actively trading (through Tradehill.) This, a freeze may not damage as many people as badly as bad as you might imagine. (But I could be wrong.) It could be disruptive if rather than freezing the BTC the government used them to actively de-stabilize global trade, but I have a hard time imagining that occurring.
I myself have hypothesized that Mt. Gox appears to receive relatively little harassment due to they being a useful monitoring chokepoint and cooperative with the authorities. While that remains in my mind a viable hypothesis, I currently consider it less likely than the null hypothesis for a variety of reasons.
Your hypothesis about keeping centralized exchanges going to reduce the pressure for de-centralized ones (and presumably private transactions and in-system utilization) to develop is also one which strikes me as entirely valid.
I currently consider the strongest hypothesis as being simply that nobody in corp/gov has really come up with a concrete plan for dealing with things. I have little doubt that the last run-up was an eye opener however, and I'll be surprised NOT to see some effort put into corralling things a bit and some observations of these efforts popping up.
I continue to believe that you may over-estimate the effect that such things will have on BTC valuations. You may even have the sign wrong.