Public policy on financial crises in emerging markets has implicitly been grounded in economic theory calling for lender-of-last-resort intervention when the country is solvent, and on theory recognizing that reputational damage is the quasi-collateral enabling lending to sovereigns with no physical collateral. The call for Private Sector Involvement — PSI — in the financing of crisis resolution has appropriately arisen from the desire for fairness as well as for successful outcomes. This paper identifies an array of PSI modalities and argues that in each crisis case the most voluntary type consistent with the circumstances should be chosen, to speed return to market access.Proper measurement of PSI is examined (e.g. gross versus net flows, time horizon). Even narrowly-defined PSI is found to have been large: $240 billion for 8 major episodes beginning with Thailand in 1997. Encouragingly, of the total fully half has been voluntary (market-based maturity-stretching swaps in Argentina and Turkey) or quasi-voluntary (bank credit line maintenance in Brazil, short-term debt conversion in Korea). A broader definition of PSI incorporating voluntary reflows after the crisis shows even larger magnitudes. The implication is that PSI has been working and does not need to be overhauled on a mandatory basis.The paper cites the growing empirical literature rejecting the hypothesis of strong "moral hazard" from the large official support programs, which is consistent with the common sense observation that lending to emerging markets has been drying up rather than mushrooming. On the proposed Sovereign Debt Restructuring Mechanism, the discussion notes the inherent difficulty of avoiding a signal that default will be facilitated for international political purposes, with resulting adverse effects on capital flows. The paper reviews the major recent crises (Argentina, Brazil, Turkey). It rejects an influential critique that the Argentine mega-swap and final IMF program in 2001 were serious mistakes, and agrees with those who judge that Brazil can sustain its debt if President-elect Lula adheres to his pledge of fiscal prudence. The analysis also examines the notion that sovereigns "wait too long" to default, and finds that the conditions for this argument to hold are stringent: the escalation of default damage from delay must be severe to offset even a moderate probability that default damage can eventually be avoided altogether. Correspondingly, the IMF should lean toward "Type I" errors of lending even when solvency is uncertain, rather than the reverse "Type II" errors, especially in light of its preferred-creditor status. At the same time, the solvency diagnosis should be informed by the growing experience showing political coherence is the most important condition for solvency, as political fragility or collapse precipitated the most serious instances of default (Russia, Indonesia, Argentina).