Analytical literature on the psi debate
Parallel to the policy debate, some authors have analyzed prospects of PSI from a more academic point of view. In essence, Eichengreen and Ruhl (2001) argue that official announcements to forcefully bail-in the private sector are not credible and will thus not affect investor behavior. Tillmann (2005) and McBrady and Seasholes (2000) oppose this view. In their empirical contributions, they show that the market actually reacts to PSI policy announcements through changes in prices and risk perceptions. Quite different is the argument put forth by Lipworth and Nystedt (2001) who use a theoretical model to show that past experiences are generally not well suited to evaluate the future efficacy of PSI initiatives. The authors emphasize the role of forward looking adaptation of market participants.
Surprisingly, none of these more analytical articles clearly and explicitly defines PSI or bail-in—even though these terms are deliberately used and at the core of the research question. Apparently, given the lack of a clear-cut general consensus, the authors have implicitly adopted the official sector’s “accepted jargon” more functional attempts for a definition of psi
Only a few authors have proposed helpful attempts to define the concept of PSI. In an important contribution, William Cline (2004) classifies modes of PSI as a function of the degree of private creditor willingness to engage in crisis resolution. Cline takes a private creditor perspective, opposed to the earlier official sector view, and puts forth three main categories of PSI: spontaneous, quasi-voluntary, and involuntary PSI. He then goes on and attributes a series of past crises and their resolution instruments to each of these categories.5 As Leiderman emphasizes (2004), Cline’s most valuable contribution is probably that he discusses a series of measurement issues and provides first rough estimates on the size of PSI in past crises.
In a similar vein, Roubini (2004, 101-2) categorizes PSI on a spectrum of voluntary and more involuntary types of private sector burden sharing. He states that defaults such as in Argentina, Russia, or Ecuador should be regarded as very coercive PSI, while cases with large bailouts (Mexico in 1995) or semivoluntary rollover agreements (Brazil in 1999, Turkey in 2001) were “softer” types of PSI. Similarly, Roubini and Setser (2004, 373) categorize PSI and ways to enforce private crisis financing into “voluntary and catalytic means,” “semi coercive” steps, and “fully coercive” steps. In their view, PSI is essentially equal to the volume of debt that was rolled over or restructured (148). No further differentiation or discussion on measuring PSI is provided.
The focus on private sector efforts and financial burden sharing is also reflected in a comprehensive study by Thiemann et al. (2005, 6) at the European Central Bank (ECB) who define PSI as “the contributions or efforts of private sector creditors to the crisis resolution process.” Note that as it stands, this definition is much more general than the earlier one. In their analysis, however, Thiemann et al. mainly limit private sector involvement to financial burden sharing.
A review of the policy debate on PSI shows that the term is not easy to define and actual crises are difficult to categorize. The official sector view of PSI as a “policy goal” is ill suited for systematic analysis and in a bulk of contributions and reports PSI remains a catch-all category, avoiding the politically delicate concrete specification of types of instruments or contributions that would count as private sector involvement or bail-in.
Cline (2004) and Roubini (2004) provide more specific categorizations. However, the categories proposed lack a clear systematic foundation. For example, Cline’s dividing line between “voluntary” and “quasi-voluntary” remains blurry. His categories are built inductively, based on his own extensive knowledge and judgement of past debt crises. He does not offer objective criteria to distinguish and delimit the categories and subcategories he presents.
Summarizing, there is no convincing definition of PSI, nor an appropriate categorization of crisis cases. Also, there are only very few contributions that do provide a functional definition of “PSI” or “bail in.” Most often, PSI or bail-in and burden sharing are simply used as a self-explanatory buzzword.
However, the concept is much less self-explanatory than it might seem at first. In fact the perspectives on, and the understanding of, PSI differ considerably between authors. It is remarkable that a standard publication in this field simply postulates that no consensus on the concept of private sector involvement and commitment in debt crises appears to exist (Roubini and Setser 2004, 140). Our first goal in this paper is thus to develop a more concise definition of PSI and its subcategories.