Broad-Based Asset Management Programs

As outlined in the Background section, high levels of NPAs can have significant negative consequences for financial institutions. It is thus not surprising that the BBAM programs studied were typically not introduced in isolation, but rather alongside other measures intended to address some of the consequences that can result from NPAs including the loss of liquidity and the erosion of capital. Interventions such as emergency lending facilities, asset guarantees, liability guarantees, corporate restructuring initiatives, recapitalizations, banking sector restructuring, and bank regulatory reform are common counterparts to BBAM initiatives. Table 3 below shows which of the programs studied were combined with these other interventions.

The decision about what to include in the package follows directly from the diagnosis of the problem, particularly the extent and type of NPAs. A country with a relatively small NPA problem focused on commercial real estate might require only a voluntary BBAM program, tied to voluntary recapitalization. A country with a larger NPA problem focused on corporate loans might also need a major corporate restructuring initiative. A country whose banking troubles are focused on a small number of systemically important institutions might be better served by company-specific rescue plans and a BBAM program only for certain types of commonly held assets, which may lend themselves to economies of scale.

Of the potential measures that could join BBAM programs as part of a broader bank support package, recapitalization is often the single most important counterpart. Banks with high levels of NPAs, almost by definition, face capital shortfalls. The exact need for recapitalization is closely tied to a BBAM program’s approach to pricing. As discussed in more detail in the section on Pricing, policymakers must decide whether to acquire NPAs at book value or for some lower amount, such as real economic value or market price. The former approach protects transferring institutions from losses on the sale of NPAs and shifts those losses to the BBAM program. This acts as a recapitalization of the transferring institutions and can eliminate the need for a separate capital injection program (but can also have significant drawbacks as will be discussed in the Pricing section).

Acquiring assets for some lower amount, by contrast, imposes losses on the transferring institutions. Depending on the size of those losses, an institution’s capital may become so eroded as to require recapitalization. The failure to pair a BBAM program with a recapitalization measure under these circumstances would likely prevent policymakers from achieving their desired outcomes. As will be discussed in more detail in the Evaluation section, an analysis of asset segregation efforts in Europe from 2000 to 2016 suggests that the removal of NPAs without recapitalization does not improve loan growth or reduce future NPA levels. The study finds that the reverse is also true—that recapitalization without removal of NPAs is ineffective (Brei et al. 2020). The BBAM programs studied here likewise provide evidence of a need to combine NPA removal with recapitalization. Hungary’s Loan Consolidation Program (LCPs), launched in 1992, failed to sufficiently reduce NPA levels by itself, resulting in the introduction of a recapitalization measure in 1993–1994. Kazakhstan’s Rehabilitation Bank similarly began acquiring NPAs without capital injections. The country’s largest banks remained under significant financial stress and ultimately had to be rescued.

The positive interplay between BBAM programs and recapitalization measures may mean that eligibility for one should be contingent on participation in the other. Several of the programs studied adopted this approach. For Mexico’s Fondo Bancario de Protección al Ahorro (FOBRAPROA), recapitalization was not only required for participation in the BBAM program, but it also determined the amount of NPAs to be acquired from a given institution—approximately two times the amount of new capital it received.^FIn the case of FOBAPROA, most of this capital was provided by private sources with government encouragement rather than coming in the form of a direct recapitalization by the state.

For efforts combining BBAM programs with other measures such as recapitalizations, the next important consideration is whether the initiatives should be combined under a single entity/umbrella or split up among different entities/umbrellas, each with a singular focus. The latter approach appears to have been more common among the programs studied. Malaysia, for example, established Pengurusan Danaharta Nasional Berhad (Danaharta) to remove NPAs from the system and Danamodal Nasional Berhad to recapitalize banks. As will be discussed in Mandates below, a combined approach may give rise to concerns about conflicting mandates and a loss of focus.

The removal of troubled assets from the balance sheets of financial institutions is not typically something that is within the ongoing mandates of central banks or other governmental bodies. Thus, the majority of the programs studied required new laws in order to be established. One unique exception to this general approach was Hungary’s Magyar Reorganizációs és Követeléskezelö (MARK), which the Hungarian central bank created pursuant to its macroprudential policy mandate.^FThe use of an existing authority such as a macroprudential policy mandate to establish a BBAM program may allow for more rapid introduction of a program during a crisis in which speed is often of the essence, but should be weighed against other potential considerations such as the need for the political legitimacy that can result from specific legislative authorization of a program. In Thailand, policymakers made use of a royal decree to expedite establishment of the Thai Asset Management Company (TAMC). Uruguay’s central bank appears to have relied on unclear legal authority to launch the Nonperforming Portfolio Purchase Scheme (NPPS) initially before it was ultimately ratified through new legislation. This resulted in several investigations by Parliament, the comptroller, and the court system. The typical need to pass new legislation authorizing BBAM programs and the added time and difficulty this can involve especially in crisis conditions may mean that it is worth considering proactively establishing the legal basis for programs to be created under certain circumstances. This could be done before the next crisis hits.

An additional consideration for BBAM programs established in the European Union (EU) has been the need to comply with European Commission (EC) rules on State aid.^FThe legal requirements for BBAM programs adopted in the EU have continued to evolve since the time of the programs studied in this survey with the introduction of the Bank Recovery and Resolution Directive and the Single Resolution Mechanism Regulation. For a complete discussion of these requirements, see EC 2018. These rules significantly influenced program design, especially with respect to the price to be paid for NPAs, as discussed in greater detail in the Pricing section.

As noted in the Overview, the legal context in which a BBAM program operates significantly influences its likelihood for success. An evaluation of the robustness of its legal framework is an important step for any country contemplating a BBAM, with particular attention to whether insolvency, bankruptcy, and foreclosure laws are strong enough to allow creditors to efficiently pursue recoveries. Sometimes this evaluation will result in the conclusion that the existing framework is sufficient, as was the case with Sweden’s Securum/Retriva, where policymakers explicitly determined that special powers were unnecessary. However, oftentimes countries introducing BBAM programs lacked effective legal frameworks for dealing with NPAs. This is often the case for emerging markets and developing economies. It appears to have been particularly true for countries whose BBAM programs were introduced as part of making the shift toward market-based financial systems. One way in which policymakers sought to overcome deficiencies in their existing frameworks was to equip BBAM programs with special powers. BBAM programs can generally be granted enhanced legal authority more quickly than a country can completely rework its existing legal framework, which may be one reason to consider adopting a program. The need to provide special powers might also weigh in favor of a BBAM program that is fully publicly owned.

Table 4 provides an overview of the various special powers included in the programs studied. As Table 4 illustrates, special powers were typically aimed at addressing deficiencies associated with inefficient or nonexistent legal processes for dealing with delinquent debtors. Often this became most important in the context of corporate debt. The problem of politically connected debtors was also a special problem in many cases.

Sometimes deficiencies requiring special powers did not reveal themselves until after the BBAM programs were already underway, making program amendments necessary. Burkina Faso’s Bureau de Recouvrement des créances du Burkina (BRCB) acquired enhanced powers to recover debt two years into its operations. Despite an initial grant of some special powers, Nigeria’s Asset Management Corporation of Nigeria (AMCON) required enhancements to its initial powers as time went on, including the ability to seize any debtor assets and the right to access debtors’ financial information.

The efficacy of these special powers appears to be mixed. The World Bank has credited Tanzania’s special tribunal and its simplified procedures and exclusive jurisdiction with having contributed to the success of the Loan and Advances Realization Trust (LART) (World Bank 2001). This success occurred despite the fact that LART dealt primarily with loans to state-controlled businesses, which, as discussed in Eligible Assets, can be a challenging asset class to address via a BBAM program. In Malaysia, Danaharta’s special legal authority enabled it to reduce the time needed to restructure loans from 14 to 16 months to two to three months. However, special powers can only work if people are willing and able to use them, which was not always the case. Indonesia’s Indonesian Bank Restructuring Agency (IBRA) possessed powerful enforcement mechanisms but was (1) generally unwilling to go after politically connected debtors and (2) often prevented from utilizing such mechanisms by Indonesian courts opposed to their use. Mongolia established a special tribunal as part of the Mongolian Asset Realization Agency (MARA), but its usefulness was limited by an inexperienced judiciary. Evaluations of Ghana’s Nonperforming Asset Recovery Trust (NPART) are split on the effectiveness of its special tribunal, with some arguing that it helped mitigate the negative impact of a weak legal framework (World Bank 1997) and others arguing that it was too slow and debtor friendly (Klingebiel 2000). Additionally, special powers may not be enough to overcome underlying NPAs that are of especially poor quality. Burkina Faso’s BRCB recovered only 10% of the value of the assets it acquired, notwithstanding what the International Monetary Fund (IMF) characterized as its powerful tools. Special powers can also be the source for political opposition. This was the case with Thailand’s TAMC, whose special powers were criticized by opponents as being excessive. Perhaps to avoid such concerns, Malaysia’s Danaharta had an oversight committee charged with making sure that special powers were not abused.

The common mandate that all BBAM programs share is the removal of assets from the balance sheets of troubled financial institutions to incentivize new capital and allow banks to lend. In every program studied, the government was seeking to clean up banks’ balance sheets or simply taking over the assets of banks already in liquidation. However, BBAM programs often differ in what they are expected to do with those assets once they take them over. Broadly speaking, policymakers generally task BBAM programs with either rapid disposal of acquired NPAs or the more active management of such NPAs via restructuring. Considerations guiding this choice can include the minimization of taxpayer losses, the maximization of NPA value, the avoidance of further market disruption via fire sales, and the creation of new financial products.

Governments also often give BBAM programs additional mandates. Mongolia’s MARA, for example, sought to normalize the enforcement of contractual obligations as the country entered into its post-Soviet future. However, the inclusion of additional mandates beyond the main focus of a BBAM program creates the risk of conflicts or distractions. In the United States, Congress charged the RTC with maximizing the return on asset disposition efforts but also gave it additional mandates, including the promotion of affordable housing and the protection of local real estate markets. Some observers believe these conflicting mandates reduced RTC’s effectiveness, given that, for example, reserving real estate assets for low-income buyers could increase RTC’s costs while concern for local real estate markets might delay sales (FDIC 1998).

Table 5 summarizes the various mandates for the programs studied.

As noted in the Part of a Package section, one key consideration is whether a BBAM program will be combined under the same entity/umbrella with other interventions such as recapitalization. This too can result in conflicting mandates. Jamaica’s Financial Sector Adjustment Company (FINSAC) was responsible for all aspects of the country’s comprehensive effort to stabilize its banking system, including bank restructuring and recapitalization. Critics maintain that this complicated FINSAC’s operations (Escobar et al. 1999). In Slovenia, policymakers initially planned to task the Bank Asset Management Company (BAMC) with both recapitalization and the acquisition of NPAs. The Ministry of Finance subsequently decided that the government would undertake recapitalization directly, allowing BAMC to concentrate on NPAs.

An additional mandate-related consideration in some cases was whether or not the BBAM program should be responsible for handling NPAs associated with particular, politically powerful debtors. One risk of having a BBAM program handle such NPAs is that the program could become a bailout for well-connected debtors or at least be perceived as such. As a distinct subcomponent of its overall operations, South Korea’s Korea Asset Management Corporation (KAMCO) acquired the debt of the failed Daewoo Group, at the time one of the country’s largest and most connected conglomerates. Critics have argued that KAMCO’s purchase of Daewoo debt was politically motivated and contributed to the program’s underperformance when the debt proved difficult to resolve (He 2004). In contrast, policymakers excluded the politically sensitive Rumasa conglomerate from the asset management activities of Spain’s Deposit Guarantee Fund (DGF), arguing that including it would go beyond the scope of the Fund’s mission. This decision has been widely praised by international observers (Sheng 1996).

In announcing the establishment of a BBAM program, policymakers face a number of potential risks related to how the program will be perceived. Acknowledging the existence of a significant NPA problem could worsen panic in the financial system. Rightly or wrongly, the program may be seen by some as a bailout for banks (as was the case with Ireland’s National Asset Management Agency [NAMA]) or politically connected debtors (as was the case with Indonesia’s IBRA). These risks and the ongoing need to maintain popular support highlight the importance of developing an effective communication strategy as part of a BBAM program.

Successful BBAM programs were typically transparent with the public about their goals and their progress in meeting those goals. Experts and international organizations have praised BBAM programs such as Sweden’s Securum/Retriva, Malaysia’s Danaharta, and South Korea’s KAMCO for their transparency and the confidence this instilled in the public. Examples of programs that have been criticized for their lack of transparency include Indonesia’s IBRA, Nigeria’s AMCON, and Thailand’s TAMC. Appropriate accounting and data standards can help improve transparency by ensuring that information is provided in a way that can be understood and trusted (EC 2018).

For some of the programs we studied, more specific information is available about their communications strategy. One key message appears to have been that the removal and management of NPAs was necessary for the appropriate functioning of the financial system and economy. Perhaps most notably, Thailand’s TAMC featured very prominently in the 2001 election for prime minister, with the winning candidate campaigning on the message that TAMC was necessary to accelerate the country’s economic recovery. Depositor protection was a key theme of messaging for the RTC in the United States and the Deposit Guarantee Fund in Spain. Given its focus on consumer assets, the United Kingdom Asset Resolution (UKAR) in the United Kingdom emphasized the consumer-friendly nature of its practices.

In Indonesia, IBRA’s initial attempt to restrict communications about its activities resulted in continued negative market sentiment and considerable criticism of the program by detractors who mistakenly believed that IBRA was non-operational. It may also have interfered with the work of IBRA officials onsite with targeted banks.

BBAM programs take on NPAs in crisis conditions, often where no market for such assets exists and with the risk of significant further losses. Private participation in the ownership of BBAM programs can thus be difficult to secure. There can be additional, affirmative reasons for government ownership of BBAM programs. Public ownership of a BBAM program may make it more feasible to invest the program with special legal powers. Additionally, the absence of private interests may more easily allow for considerations beyond maximizing return (although as discussed in the Mandate section, drifting too far from a commercial purpose may hinder BBAM efficacy). Whether for some or all of these reasons, the vast majority of BBAM programs studied were publicly owned.

For the minority of programs that involved at least some private ownership, the reasons appear to have been largely related to legal and accounting considerations. Ireland’s NAMA was 51% privately owned in response to then-current guidance from Eurostat^FEurostat is the statistical office of the European Union and “ensures the production of European statistics according to established rules and statistical principles.” that asset management companies that were majority privately owned would be kept off the government’s balance sheet. Spain’s La Sociedad de Gestión de Activos Procedentes de la Reestructuración Bancaria (SAREB) had a 55% private ownership majority for the same reason. Slovenia’s BAMC had hoped to adopt a similar approach, but the state-dominated nature of the Slovenian banking sector made private participation difficult to secure and BAMC ended up being fully publicly owned.^FThis highlighting is not intended as an assertion that private ownership of BBAM vehicles is a promising approach. Instead, it represents the view that if private ownership is pursued, it may be difficult to actually achieve in a banking system dominated by the state.

Despite majority private participation, ownership can be structured so as to leave the state with most of the control. NAMA, for instance, was largely government controlled despite the 51% private ownership stake.^FThis highlighting is not intended as an assertion that private ownership of BBAM vehicles is a promising approach. Instead, it represents the view that if private ownership is pursued, mechanisms for retaining government control over the vehicle may be an important consideration. For more details on how NAMA accomplished this control despite private ownership, please consult the underlying NAMA case study.

The legal considerations surrounding private participation in the ownerships of BBAM programs in the European Union are continuing to evolve. It is important to note that notwithstanding Eurostat’s guidance, credit rating agencies still considered BBAM program debt to be on government balance sheets, which calls into question the usefulness of this approach. In any event, the approach has been rendered moot by changes to Eurostat’s guidance eliminating the ability to keep BBAM debt off government balance sheets via private ownership participation. For a complete discussion of the legal considerations surrounding BBAM programs in the EU, see EC 2018.

Most instances of government-owned BBAM programs involved the creation or repurposing of a special entity to acquire and manage the NPAs. A minority of programs omitted this step. Under the Uruguayan NPPS, the Central Bank of Uruguay acquired NPAs directly onto its balance sheet. In Hungary’s LCPs, the Ministry of Finance acquired NPAs before transferring only a portion of them to the Hungarian Investment and Development Bank. In Mongolia, policymakers initially contemplated establishing MARA as a separately incorporated public enterprise to make it less political before deciding that it should be a government agency because they thought giving MARA governmental standing would help promote stability in the banking sector. Six years after its establishment, MARA ultimately became a public enterprise.^FDespite representing a change to the program, this feature has not been highlighted in yellow based on our determination that the change occurred due to evolving circumstances rather than a belief by policymakers that the original design was ineffective.

A fundamental challenge of BBAM program governance is that programs must be institutionally independent given the potential for political pressure, but must also be subject to mechanisms for holding them accountable (Cas and Peresa 2016). Governance of BBAM programs must thus encompass both the internal workings of programs and their external oversight.

The use of an appropriately structured Boards of Directors to oversee BBAM programs is a key element of BBAM program governance (Ingves, Seelig, and He 2004). Important considerations include the mix between public and private sector representatives, the potential role of international experts, and the independence of directors.

Table 6 below summarizes the approaches to Board composition taken by programs studied.

Certain programs studied were also subject to external oversight bodies, but with mixed results. In the United States, for example, the RTC had to appear before Congress semiannually to answer questions about its activities, while also submitting audited financial statements and regular reports. Congress also mandated that the RTC submit to analysis by the Government Accounting Office (GAO). Early criticism from the GAO resulted in substantial changes to certain aspects of the RTC’s operations over time, with the GAO ultimately praising the RTC for having made significant improvements in response to its oversight (GAO 1995).

In Jamaica, by contrast, a Commission of Enquiry established by Parliament to oversee FINSAC’s debt recoveries failed to even produce a report and was ultimately denied additional funding.

Among the most troubled of the BBAM programs studied from a governance standpoint was Indonesia’s IBRA. Despite being nominally independent, IBRA faced repeated instances of interference including the termination of its head by President Suharto. IBRA’s activities were also marred by allegations of widespread corruption, resulting in lawsuits and even criminal charges.

The programs studied did not generally adopt specific limitations on the amount of NPAs that could be acquired, although often they were established with some expectation for what the size would ultimately be. Those programs that did specifically restrict size did so indirectly, through caps on the amount of debt that could be issued in support of the programs’ operations. Like many BBAM programs, Ireland’s NAMA acquired NPAs in exchange for debt securities, but it was limited to issuing up to €54 billion in such securities before needing to seek further authorization from the government. Slovenia’s BAMC similarly acquired NPAs in exchange for debt securities, with a €4 billion cap on such issuances.

BBAM programs must generally seek the middle-ground approach of being neither too small nor too large given their specific contexts. A program that is too small may fail to adequately address the banking system’s NPA problem or to justify the effort and expense of a centralized solution. A program that is too large imposes a number of challenges, including increased operational difficulties and heightened risk to the taxpayer for programs that are publicly funded. As further discussed in the Eligible Assets section, although Malaysia’s Danaharta did not set an explicit program size, it determined an amount of NPAs that it thought was manageable and then established asset size thresholds based on an analysis of how many NPAs it would end up acquiring based on those thresholds.

Consistent with the fact that the majority of BBAM programs studied were fully publicly owned, most of the programs studied were also fully publicly funded. This public funding could come directly from the taxpayer, be raised with the issuance of government bonds, take the form of government-guaranteed bonds issued to the transferring institutions as consideration, or be paid for by the central bank (sometimes with the government later taking on the central bank’s obligations). However, even some publicly owned programs incorporated features intended to share losses with private parties. Funding programs through assessments on participating banks or the financial system more broadly might provide important political and communication leverage because those viewed as responsible for the problem are seen as having to pay for addressing it, but this must be weighed against the risks associated with placing an additional financial burden on the very institutions that are weak and in need of rescue.  

Ireland’s NAMA, in addition to being majority privately owned, included several mechanisms for sharing losses. If NAMA ended up sustaining losses in the aggregate, it had the ability to tax participants to be made whole. The consideration NAMA paid for acquired assets was also 5% comprised of subordinated debt issued by NAMA and tied to its financial performance (with the remaining 95% in government-guaranteed bonds). Several programs sought to relieve the financial burden on the government over time through broad, ongoing assessments on the financial sector. Nigeria established a Banking Sector Resolution Cost Fund (RCF) alongside publicly owned AMCON to absorb the losses on acquired NPAs. The RCF was funded over a 10-year period with contributions from the Central Bank of Nigeria (33% of RCF funding) and from annual 30-to-50-basis-point assessments on the assets of all institutions eligible for AMCON (67% of RCF funding). Mexico funded publicly owned FOBAPROA with 30-to-70-basis-point assessments on the liabilities of banks. Spain funded the publicly owned Deposit Guarantee Fund with equal contributions from the Bank of Spain and the Spanish banking sector, with banks subject to a 10-to-30-basis-point assessment on deposits. Such efforts were not always enough to prevent further losses to the taxpayer. The RCF, for example, proved insufficient to ultimately cover all of the losses associated with Nigeria’s combined BBAM and recapitalization initiative.

And as will be discussed in more detail in the section on Pricing, multiple programs studied contained ex post purchase price adjustments intended to minimize losses in the first place.

Inadequate and/or uncertain funding can undermine BBAM programs by interfering with their ability to conduct operations and reducing their willingness to dispose of assets that will impose additional losses. The need to go back for additional funding may also increase the risk of a loss of political independence. Analysts have cited funding problems as undermining effectiveness in evaluations of programs including the United States’ RTC (FDIC 1998) and Mexico’s FOBAPROA (Klingebiel 2000). Hungary’s proposal to temporarily fund asset purchases with bridge loans from the Hungarian central bank violated the European Central Bank’s prohibition on monetary financing and prevented Hungary’s MARK from conducting operations.

As illustrated in Table 1 in the Overview, the programs studied include a mix of open-bank programs, closed-bank programs, and programs for both open banks and closed banks. This determination would typically follow from a diagnosis of the situation through a stress test, asset quality review, or some other method. Having ascertained if there are banks needing to be liquidated, banks needing support, or some combination of both, policymakers can determine if a BBAM program is appropriate and, if so, define eligibility accordingly.

Programs were generally available to banks in the relevant jurisdiction (or to failed banks in the case of closed-bank programs), with some exceptions. Even some open-bank programs limited participation to institutions that met certain criteria for distress, including falling below required capital adequacy ratios (as was the case with the bank-oriented component of Hungary’s LCPs) or failing stress tests (as was the case with Spain’s SAREB). Other open-bank programs, such as BAMC in Slovenia and China’s 1999 AMC program, targeted the largest financial institutions in the country.

Whether or not participating institutions are state-owned can be an important consideration, with implications for other design features, such as the valuation of the acquired assets as discussed in Pricing. Given the widespread use of BBAM programs in countries transitioning away from state-controlled financial systems, several programs studied involved significant participation by state-owned banks. In Tanzania, LART initially limited participation to state-owned institutions, before expanding eligibility to private institutions after the program was underway.^FDespite representing a change to the program, this feature has not been highlighted in yellow based on our determination that the change occurred due to evolving circumstances rather than a belief by policymakers that the original design was ineffective.

Nonbank financial institutions were included in a minority of programs studied. South Korea’s KAMCO allowed nonbank financial institutions, including investment trust companies, insurance companies, and securities firms, to participate, although it ended up acquiring most assets from banks. The definition of eligible financial institutions under Japan’s Resolution and Collection Corporation (RCC) encompassed certain specialized institutions, such as agricultural and fishery cooperatives. Thailand’s TAMC included state-owned and private AMCs because the country initially attempted a decentralized approach to dealing with its NPA problem before deciding that a centralized AMC was necessary.

A key consideration in defining eligible institutions is whether to make participation in a BBAM program mandatory or voluntary. For the programs studied that involved a closed-bank component, participation was generally automatic upon an institution’s reaching a given state of failure. For open-bank programs, participation was sometimes mandatory for institutions meeting certain requirements.

For voluntary programs, the price paid for acquired NPAs is obviously a primary driver of participation as is discussed in Pricing. Among the programs studied, policymakers sometimes also included incentives for participation. One common incentive was requiring banks seeking to take advantage of the concurrent recapitalization program to participate in the BBAM program. This was the case for Malaysia’s Danaharta, which also had a multi-faceted “carrot and stick” approach to encourage eligible institutions to participate. In addition to a profit-sharing system discussed in more detail in Pricing, Danaharta allowed institutions transferring NPAs to amortize the losses over a five-year period. Conversely, any NPAs not transferred had to be written down immediately. Institutions also had only one opportunity to transfer a given NPA. Additional tools used in other programs studied include capital surcharges on targeted assets remaining on bank balance sheets (Hungary’s MARK) and caps on NPA levels (Nigeria’s AMCON).

Some programs approached eligibility by targeting debtors rather than creditors. In addition to a bank-oriented component that took the traditional approach of acquiring NPAs from targeted banks, Hungary’s LCPs had a firm-oriented component that acquired NPAs made to select “strategic” state-owned enterprises regardless of the originating bank. Kazakhstan’s Rehabilitation Bank focused on the country’s largest delinquent debtors.

As noted in the Background section, the types of eligible assets can determine the success of a program, with certain types of assets being particularly ill advised. The assets eligible to be included must thus be defined with particular care. Among the asset types that may be a poor fit for BBAM programs are:

• Loans to state-owned enterprises or other politically connected debtors, which can increase the likelihood of government interference, as was the case in several programs studied including IBRA in Indonesia, Société Nationale de Recouvrement (SNR) in Senegal, and FOBAPROA in Mexico.
• Residential mortgages, which can also be politically sensitive, as was the case in Jamaica’s FINSAC.
• Consumer loans and loans to small- and medium-sized enterprises, which may require too much direct knowledge of the debtors.

Assets typically better suited to BBAM programs include commercial real estate and large corporate exposures, which are generally larger in size, less politically fraught, and backed by collateral that is easier to realize upon.

Additional complications can arise if the assets acquired are too heterogeneous or include too many small exposures to be effectively managed. Finland’s Arsenal accepted NPAs of all types and sizes, a fact that evaluations of the program have cited as hindering its efforts to dispose of assets (Klingebiel 2000). To address the concern about the efficiency of handling small assets, several programs studied included minimum asset size requirements including Securum/Retriva in Sweden, Danaharta in Malaysia, and SAREB in Spain. In Malaysia, policymakers established the minimum size threshold at the point at which they thought they could manage the workload, having analyzed what loans that meant they would be buying.

Certain programs studied went beyond simply excluding small exposures and specifically targeted NPAs from the most significant delinquent debtors in the country. This was the approach taken by Kazakhstan’s Rehabilitation Bank and the firm-oriented component of Hungary’s LCPs. The ability to aggregate the debt of such large debtors and thereby increase negotiating power is one potential benefit of a BBAM approach.

One of the potential concerns associated with BBAM programs is the risk of moral hazard. If banks have an expectation that they will be able to transfer NPAs to a BBAM program, they may be more likely to extend risky loans. Several programs studied sought to address moral hazard by limiting eligibility to assets initiated by a given date. China’s 1999 AMC program accepted only NPAs generated by policy lending prior to 1996. The bank-oriented component of Hungary’s 1992 LCPs excluded loans made after October 1, 1992. Tanzania’s LART initially restricted eligibility to assets in existence at the time of the legislation establishing the program.

The risk associated with being too restrictive in the assets accepted by a BBAM program is that it may leave banks with NPA levels that remain problematically high. This appears to have been the case with Thailand’s TAMC, in which different sets of asset eligibility criteria were developed for state-owned institutions and for private institutions. The criteria for private institutions were more stringent, with NPAs subject to a previous court ruling or restructuring agreements deemed ineligible. Given that Thailand had already been in crisis for several years by the time of the TAMC’s introduction, this excluded a majority of NPAs held by private institutions. As a result, the TAMC acquired the vast majority of its assets from state-owned institutions, while private banks and AMCs retained high levels of NPAs.

While the nature of BBAM programs is such that they typically acquired NPAs specifically, many programs, such as the RTC in the United States, the RCC in Japan, and the BAMC in Slovenia, also acquired performing loans. This may have been motivated by the belief that consolidating all of the obligations of a given debtor, whether performing or not, would give the BBAM program additional leverage and promote more effective restructuring. This was the motivation explicitly cited by the BAMC. It may also be intended to help the BBAM program achieve stand-alone profitability. However, there may be circumstances in which specific performing assets are included for political reasons. Nigeria’s AMCON purchased performing loans that it deemed too large relative to the capital base of the transferring banks, but critics allege that these assets were included because a government official was the debtor (Cerutti and Neyens 2016).

Policymakers designing BBAM programs must determine the mechanics of acquiring the targeted assets. The approaches here are generally very different for closed-bank programs and open-bank programs. Closed-bank programs typically acquire assets automatically upon an institution’s failure, often with no need to make decisions about which specific assets to acquire and what consideration to provide. In some instances, negotiation over which NPAs a closed-bank program would take on were a key feature of attempts to sell failed banks to acquiring institutions, as was the case for the United States’ RTC and the Spain’s Deposit Guarantee Fund.

Open-bank programs, conversely, must choose which NPAs to acquire, as well as the timing, logistics, and consideration to be used. In many cases, participating institutions provided BBAM programs with lists of eligible assets that the programs would then evaluate for acquisition. In Hungary, MARK selected the specific assets to be acquired at random off a list of all eligible assets to avoid adverse selection and help ensure that assets were acquired at an appropriate price.^FWhile the random selection of assets off a submitted list to reduce adverse selection was a key feature highlighted by the European Commission in its consideration of Hungary’s MARK, the ability of random selection to effectively combat adverse selection is dependent on the submitted list itself not being the product of adverse selection. Randomly selecting assets off a list of an institution’s worst NPAs would still leave a BBAM program with the worst NPAs. Such a design feature is thus interesting, but also to be approached with caution.

The timing of asset acquisition by open-bank BBAM programs can create moral hazard concerns. Ongoing asset purchases may encourage banks to engage in risky lending with the expectation that any NPAs that result can be transferred to the program. For programs that want or need to make ongoing purchases, this can be addressed by limiting the assets that can be transferred to those existing as of a certain date, as discussed in Eligible Assets. Several programs did in fact employ a tranche approach, in which assets were purchased in groups over time, in most cases seemingly to avoid the logistical difficulties that would have been associated with attempting to acquire all targeted NPAs at once. Malaysia’s Danaharta acquired NPAs in two rounds—an initial primary carveout and a secondary carveout with more selective criteria. Thailand’s TAMC sorted NPAs into tranches based on the number of creditors, institution type, and amount of outstanding debt. Slovenia’s BAMC divided purchases among multiple tranches, with all of the assets associated with a given debtor included in the same tranche. Spain used the health of the transferring financial institutions to split SAREB’s acquisition into two groups, with the most distressed institutions transferring first. In Ireland, NAMA had to abandon its initial plan of sorting assets into tranches where the first tranche contained a small number of debtors with large exposures and later tranches would include more debtors with smaller exposures. This multi-step process was proving to be too slow. The need to accelerate the pace of asset acquisition resulted in the collapsing of multiple anticipated future tranches into one final bulk tranche.

The consideration paid for acquired NPAs in open-bank programs typically consisted of debt securities, often guaranteed by the government. The use of debt securities avoids the need to provide cash up front and instead allows for payment over time as the program begins to realize returns on acquired NPAs. This is a concern primarily for budget-constrained governments that would otherwise not have the ability to fund the purchases. In such situations, bond maturity should be tied to the lifespan of the program and the expected timing of cash flows from the NPAs (Ingves, Seelig, and He 2004). Another approach is to use zero-coupon bonds to delay the need to make any payments until bond maturity, as was done in Nigeria’s AMCON. If the bonds are eligible collateral for central bank lending facilities, they can also be used to improve recipient banks’ liquidity position right away. A less common form of consideration seen in Kazakhstan’s Rehabilitation Bank and Senegal’s SNR was the removal of NPAs from bank balance sheets in exchange for the cancellation of a corresponding amount of bank liabilities owed to the government. These liabilities could take the form of previous government assistance and/or deposits from state-owned enterprises or other agencies of the government.^FThe ability to net the purchase price for transferred assets against amounts already owed by the transferring institutions to the state may be a way of simplifying the process of acquiring assets in some circumstances.

A key consideration for the design of a BBAM program, particularly one that will engage in open-bank acquisitions, is the price at which assets will be acquired. Cerruti and Neyens (2016) have described this as the most difficult of all of the design decisions associated with establishing a BBAM program. As discussed throughout this survey, it is also a decision that has significant implications for other program features.

As illustrated in Table 7, approaches to pricing can generally be divided into one of three categories: acquisition at book value, acquisition at real economic value, or acquisition at market price. In crisis conditions, it is generally understood that acquisitions at both book value and market price are likely to result in pricing that is different from the actual value of the asset. If an NPA is acquired at book value, the price will be higher than the asset’s ultimate return, given its nonperforming status. This shifts the losses associated with the NPA from the transferring institutions to the BBAM program. If an NPA is acquired at market price, the price will be lower than the asset’s ultimate return, given the market-distorting effect of the crisis.

Both paying too much and paying too little for NPAs have the potential to undermine the effectiveness of BBAM programs. Paying too much may obviate the need for recapitalizing the transferring banks as discussed in the Part of a Package section, but there are potential concerns associated with this form of indirect recapitalization. These include (1) the fact that capital provided in this form is free or low cost and therefore disadvantages other firms and (2) the fact that direct injections typically afford policymakers more control over the institutions receiving the capital (Rose 2005). Paying too much may also undermine attempts to deal with the NPAs themselves. If BBAM programs acquire NPAs at prices above what can be realized from the assets, the disposal of the NPAs will result in large, immediate losses. The desire to avoid such an outcome may result in BBAM programs’ holding on to NPAs without any real effort to dispose of them. If acquiring NPAs at higher values is necessary despite this concern, an immediate revaluation on the books of the BBAM program may be important in removing any disincentive to disposal (Rose 2005).

As indicated in Table 7, acquisition at book value has been a frequently used approach, notwithstanding the concerns associated with paying too much for NPAs. Sometimes this may be a practical consideration. The Spanish Deposit Guarantee Fund, for example, purchased NPAs at book value because policymakers considered it too difficult to determine an alternate, market-based price. It is also the case that book value pricing may be less of a concern where the participants in a BBAM program are state-owned banks, because the state will be responsible for the losses whether they are imposed on the program or on the transferring banks.

Paying too little for NPAs introduces its own set of potential challenges. If institutions perceive the price offered for NPAs as too low and participation is voluntary, there may be little uptake. This was initially the case for the RCC in Japan, which in its early years paid an average of 3.8% of book value and attracted few participants. Japanese policymakers later successfully amended the RCC’s approach with the explicit objective of encouraging more participation. Prices can be too low to secure participation not only in absolute terms, but also relative to institutions’ expectations about what prices will be available in future versions of the BBAM program. Hungary’s LCPs featured both an initial bank-oriented component that acquired NPAs from targeted banks and a subsequent firm-oriented component that acquired NPAs made to targeted state-owned enterprises by any banks. It appears that banks were reluctant to dispose of eligible NPAs during the initial bank-oriented component because they expected those same NPAs to fetch higher prices during the subsequent firm-oriented component.

Even if BBAM programs succeed in attracting participants at low prices (possibly because participation is mandatory), such pricing can interfere with the effective disposition of acquired NPAs. A BBAM program that acquires NPAs at very low prices may be able to achieve significant profits without maximizing the value of those assets. This could reduce incentives for maximizing value and promote hurried disposal at prices below those that might ultimately be possible.

The real economic value approach to pricing can be seen as an attempt to find the appropriate middle path between prices that are too high and prices that are too low for a BBAM program to be successful. Typically it involves an attempt to determine the intrinsic value of an asset outside of crisis conditions when the market-distorting effects of the crisis have subsided and the actual losses incurred by the asset are realized. This can prove difficult, as was the case with Spain’s SAREB, where a revaluation of assets three years after they had been acquired at what was then considered real economic value resulted in substantial losses.

For BBAM programs introduced in the European Union, purchasing assets at an above-market price constituted State aid under EC rules. Such purchases would be approved only if they met certain criteria set forth in the Communication from the Commission on the Treatment of Impaired Assets in the Community Banking Sector of February 25, 2009 (better known as the Impaired Asset Communication, or IAC) and subsequently modified in future communications. To comply with these criteria, programs generally needed to purchase assets at their “real economic value,” a figure aiming to approximate the asset’s long-term economic value based on prudent assumptions about cash flows and observable market inputs. This resulted in lengthy valuation processes and/or clashes with the European Commission that often delayed BBAM programs in the EU. For a more detailed discussion of the European Commission’s approach to pricing, see EC 2018.

Given the uncertainty and complexity associated with attempting to value assets in the midst of crisis and often with little time to do so, some BBAM programs relied on ex post purchase price adjustments based on how the NPAs ultimately performed. These could be structured to protect the BBAM program from losses resulting from overpaying, to protect the transferring banks from losses resulting from underpaying, or both. Sometimes these ex post adjustments were done only temporarily given the need to acquire an initial set of assets with considerable speed. This was the case with South Korea’s KAMCO, which aimed to make rapid initial bulk purchases, leaving inadequate time to completely evaluate assets. The AMC consequently included mechanisms for ex post price adjustments. Once market conditions stabilized, KAMCO shifted its approach to rely on more exact upfront pricing with no adjustments.^FDespite representing a change to the program, this feature has not been highlighted in yellow based on our determination that the change occurred due to evolving circumstances rather than a belief by policymakers that the original design was ineffective. In a situation in which initial large-scale asset purchases must be done quickly without adequate time for valuation, it may be worth considering making the use of purchase price adjustments temporary and transitioning to full valuation with no adjustments as soon as conditions allow.  

Other adjustments were more permanent. Perhaps the most extreme adjustment example was Indonesia’s IBRA, under which banks received no upfront consideration for the transferred NPAs but were entitled to any proceeds from their ultimate disposition. In Ireland, NAMA set an initial price for assets based on calculations of real economic value, but this amount was subject to a clawback feature that enabled NAMA to recover any amount it was determined to have overpaid on acquired assets stemming from expedited and/or inaccurate initial valuations. NAMA could also include a clawback feature to protect banks from underpayment, but it is unclear how widely used this was. Danaharta in Malaysia included a profit-sharing arrangement as an incentive for participation, whereby transferring institutions were entitled to 80% of the profits on NPA disposal. Conversely, Mexico’s FOBAPROA negotiated loss-sharing arrangements with participating banks that would leave banks responsible for 20 to 30% of the losses on the NPAs that they transferred. Under Thailand’s TAMC, transferring institutions shared in profits up to the book value of transferred assets and were responsible for losses of up to 30% of the transferring price. The TAMC and the transferring institutions calculated and recognized these profits and losses at the end of the fifth and 10th years of the TAMC’s operations.

While potentially attractive in terms of their ability to promote accurate pricing and protect BBAM programs from losses, adjustments may hinder the effective functioning of the intervention. Banks may be less willing to transfer NPAs if they remain exposed to further losses on them (Ingves, Seelig, and He 2004). And even if banks do transfer the NPAs, this continued exposure could mean that counterparties remain uncertain about the banks’ viability. Particularly where a BBAM program is intended to restore counterparty confidence in institutions to, for example, ensure their continued access to funding, ex post price adjustments may be counterproductive.

The approaches taken to the management and disposal of acquired NPAs in programs studied were closely related to their overarching mandates. An overview of select programs’ approaches is included in Table 8.

A minority approach to the management and disposal of NPAs gave responsibility for the assets to the transferring institutions, either because the institutions were better positioned to deal with them and/or because the exposures were too small to warrant the attention of BBAM program staff. Where a primary objective of a BBAM program is to free transferring banks to focus on their core business, requiring these banks to remain responsible for the management and disposition of the assets can be self-defeating. Such arrangements also require careful consideration of the banks’ incentives and contractual provisions that reflect these realities. Hungary’s LCPs adopted this approach for the bank-oriented component of its program. Mexico’s FOBAPROA also made use of the banks themselves, and as discussed in the Pricing section, negotiated loss-sharing arrangements to incentivize banks to manage assets appropriately. In Spain, SAREB initially relied on transferring banks for management and disposal before concerns about conflicts of interest ultimately prompted a shift to outside contractors.

The use of outside contractors or the sale of NPAs to third parties who will then seek to collect on them can create political challenges. In Jamaica, FINSAC transferred residual NPAs to an American financial institution whose aggressive collection efforts resulted in widespread criticism from the Jamaican public.

In some instances, attempts to build safeguards into the disposal process interfered with program effectiveness. Congress initially required the United States’ RTC to dispose of assets for at least 95% of assessed value. They later reduced this threshold to 70% to accelerate the disposition process. In Slovenia, the BAMC was subject to a requirement that it dispose of at least 10% of its assets every year, but this may have resulted in forced sales and unnecessary losses.

The need to dispose of large volumes of NPAs often resulted in the pioneering of new techniques and products. The United States’ RTC is often credited with having jumpstarted the market for commercial mortgage securitizations. Danaharta was responsible for the first collateralized loan obligation in Malaysia. The desire to develop secondary markets for NPAs and new approaches for dealing with them was often a key consideration in launching BBAM programs.

A final key consideration in the design of BBAM programs is whether or not the program should have a predefined timeframe and, if so, how long it should be. Too short a timeframe could result in a program’s disposing of NPAs prematurely in ways that exacerbate market conditions and/or fail to maximize the value of the assets. Too long a timeframe could result in NPAs’ languishing in the program and continuing to lose value over time.

As Table 9 illustrates, the programs studied took a variety of approaches to the timeframe question. Given the potential concern that too long a timeframe could result in languishing, it is perhaps surprising to see how many programs had no predefined timeframes and were effectively open ended. The IMF repeatedly criticized Nigeria’s AMCON, for example, for failure to establish a sunset date, warning that it increased financial risks and the threat of moral hazard (IMF 2013). Instead of relying on open-ended timeframes, several programs studied appear to have addressed timing uncertainty by establishing a defined lifespan subject to extension. As indicated in Table 9, several countries made use of such extensions, while others ended programs early.