I am new to the academic discussion on this topic, so I apologize for any disingenuousness that follows, but let me throw in some perspective that may or may not have been hitherto considered.

Marc Franklin told us in torts almost thirty years ago that you couldn't really understand the development of tort law in the 20th century without understanding the impact of insurance. I think the same is presently true of private securities litigation when either a corporate insider or the corporation as control person is alleged to have done something wrong. I am going to put aside for the time being the recent settlements (I am having a senior moment on which company - but I'm pretty sure it was Enron) in which the directors were forced, notwithstanding D&O insurance, to pay out of their own pockets to complete the settlement.

Almost every public company takes out directors and officers insurance. So long as a company is solvent, the insurance does nothing more than reimburse the company for the amounts it would be required to pay the accused wrongdoers under the indemnification provisions of the bylaws. For that reason, that part of the policy is referred to as "balance sheet protection." There is generally a deductible - you could expect a $1 million deductible on $50 million of coverage. So the real cost to the shareholders is not the payout to the plaintiffs, but the cost of the insurance premium. (This is also sometimes referred to as Side B coverage.)

What Os and Ds really care about is the Side A coverage: the terms of defense and indemnity when the corporation is either not around or refuses to indemnify. In that case the "nonculpable non-plaintiff" shareholders either are protected because the corporation isn't helping the bad guys, or the corporation is bankrupt, in which case they are SOL anyway. (Usually there is no deductible on this part of the coverage.)

There is also something called "pre-determined allocation". The D&O policy doesn't cover the corporation for its own liability. No plaintiffs' securities lawyer in her right mind would sue only the company on a control person theory and leave out the individuals because no insurance coverage would kick in.
Up until a couple of years ago, the policies would generally include (for a reduction in premium) this pre-determined allocation. If you had a securities case with both insured individuals and the non-insured corporation named as defendants, the parties essentially stipulated beforehand that, for purposes of Side B coverage, the responsibility would be deemed, say, 80% to the individuals and 20% to the company. In the last years I was involved in placing coverage, the insurance market was more competitive in seeking business, and you could get a 100% pre-determined allocation. Again, that meant that if the plaintiffs' bar knew what it was doing, all the liability would be attributed to the individuals and none to the corporation as control person.

That's a relatively brief explanation how the insurance works. My point is this: a debate about who bears the loss as between groups of shareholders if there were no insurance may have some theoretical impact on something somewhere, but isn't it really the case that for all intents and purposes the non-trading shareholders are protected by insurance and the whole debate is a non-issue?

Without getting into any specifics, and just to be really diabolical in wrapping various threads together, would it be a breach of the duty of care for the directors not to have taken out the balance sheet coverage (trading off current earnings for contingency protection)? Should the business judgment rule protect a director who has decided not to get insured for the innocent shareholders? To Jeremy Telman's point earlier, is this an issue of protecting the directors or the business of the corporation?

Do we know whether shareholders as a group would prefer an environment of no litigation as to an environment of accessible litigation?

Yes, we do. See Thakor's papers for ILR last year. Diversified investors break even; they would thus prefer a rule that minimizes the billions of dollars of transactions costs of securities litigation.

Certainly, New York pensionholders who got taken along for the ride by Hevesi's lawsuit against banks over WorldCom would have preferred no litigation: the pension held the stocks of both WorldCom and the defendants, and lost more from the latter than it gained as a plaintiff. For this, Hevesi's campaign contributors got $300 million in fees.

I don't read the literature as stating that injured plaintiffs receive a windfall when they are compensated. In these discussions, the term "windfall" is usually attributed to the (non-insider) shareholders who had the good fortune to sell while the price was inflated. I see the literature noting that, ex ante, a diversified investor is a winner half the time and a loser half the time, and, in the words of Easterbrook and Fischel, "perceives little good in a legal rule that forces his winning self to compensate his losing self over and over."

That still leaves the problem of non-diversified investors and inside traders, but I don't see why SEC civil (and criminal, in the case of intentional fraud) prosecution isn't sufficient for that.