Portugal: Ex-Post Evaluation of Exceptional Access under the 2011 Extended Arrangement

This paper discusses Portugal's Ex-post Evaluation of Exceptional Access Under the 2011 Extended Arrangement. Portugal faced a sudden stop in financing in 2011. The authorities' IMF-supported program aimed to address the problems that had made Portugal vulnerable to changes in market confidence. The evaluation concurs that the program's 'big decisions' were justified. The main lesson to be drawn from Portugal's experience is that adjustment in the context of currency union membership is difficult. Further work is needed to flesh out the measures required to support internal devaluation and private sector deleveraging. Options for union-level conditionality would benefit from clarification.

Abstract

This paper discusses Portugal's Ex-post Evaluation of Exceptional Access Under the 2011 Extended Arrangement. Portugal faced a sudden stop in financing in 2011. The authorities' IMF-supported program aimed to address the problems that had made Portugal vulnerable to changes in market confidence. The evaluation concurs that the program's 'big decisions' were justified. The main lesson to be drawn from Portugal's experience is that adjustment in the context of currency union membership is difficult. Further work is needed to flesh out the measures required to support internal devaluation and private sector deleveraging. Options for union-level conditionality would benefit from clarification.

This paper is an Ex-Post Evaluation of Portugal’s 2011 Extended Fund Facility (EFF) arrangement. The 3-year arrangement ran from May 2011 through June 2014, with 11 reviews completed. The authorities chose to let the arrangement expire without completing the final review. Official financing of €78 billion was provided by the Troika, in a two-to-one split between the EU (€52 billion) and the IMF, which provided €26 billion (SDR 23.7 billion, or 2,306 percent of quota)—i.e., exceptional access. The ex-post evaluation of exceptional access programs required by Fund policy: (i) reviews performance against program objectives; (ii) discusses whether program design was appropriate to address Portugal’s challenges; and (iii) assesses whether program modalities were consistent with Fund policies.1

Background to the Crisis

Portugal had the lowest per capita income of founding member countries when it joined the euro area. As in other countries on the periphery, easy financing terms associated with euro accession sparked a spending boom and build-up of debt (both public and private). But unlike elsewhere, Portugal grew slowly, with almost no income convergence. By 2010, when financing began to dry up, Portugal had twin deficits (current account and fiscal) of 10 percent of GDP, and public and private debt among the highest in the euro area.

A. The Build-up of Imbalances

1. Easy financing conditions associated with euro adoption allowed Portugal’s public and private debt to double between 1998 and 2011. Reduced credit costs funded a consumption boom and fall in domestic savings; in turn, the current account deficit and the international investment position worsened. By 2010, Portugal’s fiscal and external current account deficits were 10 percent of GDP. General government debt was 96 percent of GDP, and private sector debt was 247 percent of GDP (among the highest in the euro area), implying heavy reliance on external funding that left Portugal vulnerable to a sudden stop in capital inflows.

2. Despite easy financing, growth disappointed. Growth briefly accelerated in the wake of euro accession but faded after 2000, despite access to finance and sizeable investment from EU structural funds. Portugal made little progress on income convergence toward the EU, or even to peers such as Spain and Greece. Between 1992 and 2007, Portuguese income per capita rose only from 66 to 70 percent of the EU average (pp). The poor growth reflected limited progress in addressing longstanding structural bottlenecks, including rigid labor markets, an inefficient judicial system, obsolete insolvency legislation, and large and inefficient state enterprises. Wage growth continued to exceed productivity increases, eroding external competitiveness. Between 1995 and 2010, the ULC-based REER appreciated, bringing the competitiveness gap to an estimated 13-14 percent as Portugal lost market share.

3. Public debt rose as Portugal failed to adhere to the fiscal rules of the EU’s Stability and Growth Pact. Despite substantial savings on the interest bill, headline fiscal deficits consistently exceeded the Maastricht 3 percent of GDP deficit limit, as pensions and public wages rose. Moreover, some spending obligations were shifted outside the budget, with a loss of control over debt in local governments and autonomous bodies, and a build-up of contingent liabilities through PPPs and SOEs. Hence, public debt rose steadily following euro adoption, from 48 percent of GDP in 2000 to 61 percent in 2005 and 96 percent in 2010.

uA01fig01

Public and Private Sector Debt

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: Bank of Portugal.

4. The rise in private debt was mainly financed through domestic banks, which expanded significantly after euro adoption. Credit growth substantially outpaced domestic savings. Banks funded corporate loans and household mortgages by borrowing abroad—with wholesale funding of 40 percent of assets by end-2008. The banking system became one of the more leveraged in Europe, with the loan-to-deposit ratio rising to 160 percent by mid-2010.

5. The leverage was not put to the most productive use. Investment was disproportionately in nontradable activities, including housing (though there was no house price boom), energy, and highway infrastructure.2 Moreover, borrowing shifted from financing investment to covering sizable increases in operating costs. High debt service costs for firms (despite record-low interest rates) further eroded their profitability; their lower investment and competitiveness became a drag on growth. In hindsight, the combination of credit growth well above real GDP growth together with weak corporate profitability suggest a significant deterioration in credit underwriting standards in the decade before the crisis.

B. Crisis Triggers

6. The global financial crisis pressured both the public finances and the banks.

  • A large fiscal expansion in 2009 worsened the fragility of public finances. The fiscal deficit rose from 3.8 percent of GDP in 2008 to 9.8 percent in 2009 and 11.2 percent in 2010,3 with public debt rising by 25 percent of GDP in two years. The sharp increase in debt contributed to a ratings downgrade by S&P in April 2010 and again in April 2011.

  • In 2008, CDS spreads for the main banks rose significantly, bank funding through securities dried up, and two small banks (BPN and BPP) were intervened. By 2009, funding conditions deteriorated significantly and rating agencies downgraded the large Portuguese banks. By mid-2010 the banking sector’s access to international market funding was effectively closed.

7. Rising public debt together with the emerging crisis in Greece resulted in a widening of Portugal’s sovereign debt spreads, from early 2010. Spreads rose sharply through 2010, reaching 300 basis points by end-June and exceeding 400 basis points in early 2011—the highest level since euro adoption (Box 1).

8. The rise in sovereign yields increased financing pressure on Portuguese banks, given their heavy reliance on wholesale funding from euro area banks. The increase in sovereign risk weighed heavily on perceptions of the banking system, given concerns over sovereign-bank linkages. To offset the decline in wholesale funding, banks became increasingly reliant on Eurosystem financing (EUR 48 billion or 9 percent of total assets in early 2011). The buffer of unused repo-eligible collateral (EUR 15 billion) became insufficient given larger short-term net refinancing needs. The government announced stabilization measures and banks were able to raise EUR 3.3 billion of additional capital. Although a majority of banks had Tier 1 ratios above 8 percent, increasingly the sufficiency of buffers was questioned by market participants and Fund staff.

9. As a result, the pattern of activity of 2000-2010, with foreign inflows intermediated through local banks to finance domestic consumption, began to break down. The impact of the global financial crisis shrank exports in 2009, but growth returned in 2010 as exports bounced back to pre-crisis level. Private consumption and investment both slumped in 2011 as bank financing dried up, however, contributing to another decline in real GDP.

10. The banking system was highly vulnerable at the outset of the program, but there was no banking crisis. Banks were deemed solvent and viable, though vulnerable because of their high reliance on wholesale funds. The quality of the loan portfolio started to deteriorate sharply, but from a low base. Corporate and consumer credit were particularly affected, with NPLs rising to 5 percent and 8.5 percent respectively. But there was no public panic, bank run, or evident material losses (burst housing bubble, significant mark-downs of assets, etc.)

C. Crisis Response

11. Efforts to reverse the deterioration in public finances ran aground in 2011 as the macro-outlook worsened and market pressures mounted. The 2011 budget had targeted a large fiscal adjustment, but later it became clear that growth would fall well short of the budget assumptions. Hence the government had to include additional measures in the updated Stability Program presented to Parliament in March 2011.4 The program also included ambitious targets for further adjustment in 2012-13. It was rejected by Parliament, and the government resigned. At End-March 2011 the fiscal outturn for 2010 was released, with several public transport SOEs reclassified into general government, and a consequent increase in public debt of nearly 10 percent of GDP. The uncertainty over the political outlook and concern about debt dynamics boosted sovereign spreads further, to more than 500 basis points in early April.

12. The sharp increase in borrowing costs made the financing situation increasingly untenable. Following the resignation of the government and the call for new elections, access to external financing was effectively shut off and foreign investors’ participation in the domestic market fell sharply, leaving the government dependent on short-term domestic bank financing. With yields continuing to increase sharply, the authorities asked for financial assistance from the EFSM/EFSF and the IMF on April 7, 2011. Program negotiations began in mid-April, and an arrangement was approved by the Board on May 20.

Program Strategy and Financing

Portugal’s program was the third crisis program in stressed euro area countries. While confronting a similar fall-out from easy financing and a consumption boom, Portugal differed from Greece in having a bigger private sector debt build-up, and from Ireland because its banks were considered broadly sound. Its prolonged low growth was considered the fundamental problem, and—more so than in the two other cases—measures to support growth and strengthen competitiveness were put at the heart of the program.

A. Strategy

13. The program had three pillars, and a strong premise of market-friendly adjustment. It had a multifaceted strategy to enhance competitiveness and growth; a front-loaded fiscal consolidation aimed at sending a convincing positive signal to markets (to allay any fear that debt would be restructured); and a strategy to maintain financial stability by strengthening banks’ balance sheets without ‘excessive’ deleveraging which would undermine growth. Success on these fronts was expected to allow Portugal to return to market funding at reasonable rates within two years.

  • Enhancing competitiveness and growth. At the heart of an extensive structural reform strategy was an intended internal devaluation, to substitute for the missing exchange rate instrument by reducing domestic costs relative to trading partners’. To make exportables cheaper, reforms sought to increase labor market flexibility and cut costs of non-tradables by fostering competition in energy, transport, and telecommunications. The minimum wage was to be frozen and unemployment benefits reformed. A fiscal devaluation was envisaged, to reduce taxes paid by labor and increase taxes on consumption (making imports (and other goods) dearer). There were also measures to increase productivity, improve the business environment, and make the judicial system enforce civil and commercial claims more effectively.

  • Instilling confidence and ensuring fiscal sustainability. The program envisaged a front loaded fiscal consolidation to restore credibility and regain market access. The large size of the adjustment (about 9½ percent of GDP) was calibrated to reach a deficit of 3 percent before the end of the program, in line with Maastricht criteria. The authorities wanted even more ambitious targets, but were dissuaded by the Fund to avoid undermining growth further. Adjustment was to be split roughly 2:1 between expenditure cuts and revenue measures. Spending cuts were emphasized since expenditure had jumped in the stimulus package of 2009 and was under further pressure as social transfers grew in the crisis; moreover, spending-based adjustment was seen as more durable and less inimical to growth.5 The program also included institutional reforms to streamline the public sector and reduce fiscal risks (notably, in state enterprises and PPPs, where contingent liabilities were exceptionally large).

  • Safeguarding financial stability and avoiding excessively fast deleveraging. The program aimed to strengthen banks’ balance sheets by raising capital requirements and reducing reliance on wholesale funding. A EUR 12 billion Bank Solvency Support Facility (BSSF) was set up to backstop banks having difficulty in raising capital. To avoid too-fast deleveraging by capital-constrained banks, continued high access to ECB financing was crucial; inter alia the maximum amount of eligible government guarantees for uncovered bonds was raised to EUR 35 billion. The deleveraging, which was benchmarked against loan-to-deposit ratios (though not as program conditionality), was coordinated under the umbrella of a Solvency and Deleveraging Assessment Framework (SDAF). To guide the SDAF, banks were diagnosed in a program of special on-site inspections (SIP) and stress-testing designed by the Bank of Portugal (BdP). The program also included measures to strengthen supervision and bank resolution, and to facilitate market-based corporate and household restructuring with simplified insolvency procedures.

14. In all dimensions, the program was a balancing act between convincing upfront adjustment and growth-supporting gradualism. Fiscal measures were frontloaded, but so was access, to allow Portugal to address short-term financing pressures. Thus, even with frontloading of measures, Portugal would stabilize its debt only at the end of the program (at a targeted 115 percent of GDP). The current account deficit was projected to decline somewhat more gradually than the fiscal deficit, to below 4 percent of GDP at the end of the program—leaving room for private consumption, and reflecting the expectation that competitiveness would come slowly. The strategy for containing bank fragility did not envisage upfront interventions (as in Ireland) but relied on recurrent reviews of banks’ asset quality and the adequacy of impairment levels to ensure a realistic degree of loss recognition over time. The indicative targets for bank deleveraging were set with the goal of avoiding a credit crunch. As with most countries which are in a monetary union, the Portugal program had to rely significantly on structural conditionality (Figure 1).

Figure 1.
Figure 1.

Conditionality in Portugal’s Program and Comparators

The program had relatively few quantitative performance criteria, but higher than average structural conditionality

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: MONA Database.1/ Includes quanititative benchmarks and indicative targets.

15. It was recognized that the program entailed important risks. These included near-term government refinancing risks, the impact of lower-than-programmed growth on the debt path, the challenges of sustaining social support for fiscal and structural reforms, the need for political consensus especially as Portugal transited an election period, and the risk of realization of contingent liabilities or migration of bank losses to the sovereign. The main external risk was spillovers from the deepening problems in the euro area.

16. The program mobilized buffers against some risks, but strong prior actions were not possible. A core buffer was the capital backstop for banks (BSSF). To mitigate election-related political risks, the staff team sought support for key program parameters from all major political parties. Usually, prior actions also provide strong buffers, but, for Portugal, prior actions had to be limited to those that did not need parliamentary approval, since parliament had been dissolved.

B. Financing and Exceptional Access

17. The EFF was large by Fund standards, though it covered only one-third of Portugal’s €78 billion external financing need. European Union partners covered the rest (€52 billion). The arrangement, for SDR 23.7 billion (€26 billion, or 2,306 percent of quota), was the third largest relative to quota (after Greece and Ireland), and the joint largest relative to GDP (Figure 2). Portugal’s access during the first year was 1,117 percent of quota. The frontloading was significant but not extreme. Front-loaded resources were needed to cover immediate financing needs, including large amortizations and funding for the BSSF.

Figure 2.
Figure 2.

Financing and Access under Portugal’s EFF

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

18. For EFF approval and at each review, staff argued—and the Board agreed—that all four criteria for exceptional access were met, albeit by applying the systemic exemption clause.

  • 1. The member is experiencing or has the potential to experience exceptional balance of payments pressures on the current or capital accounts, resulting in a need for Fund financing that cannot be met within the normal limits. Both the public and private sectors faced large external financing needs, with limited access to external markets and continuing current account deficits. Private banks, facing difficulty in rolling over external debts, were already meeting some financing needs by ECB liquidity support.

  • 2. There is a high probability that the member’s public debt is sustainable in the medium term. In instances where there are significant uncertainties that make it difficult to state categorically that there is a high probability that the debt is sustainable, exceptional access would be justified if there is a high risk of international systemic spillovers (the systemic exemption).6 Given Portugal’s high debt and significant uncertainties about interest rates, growth, and contingent liabilities, staff appealed to the systemic exemption. The spillover effects were not, however, discussed.

  • 3. The member has prospects of gaining or regaining access to private capital markets within the timeframe when Fund resources are outstanding. Successful program implementation and a return to growth were expected to reduce Portugal’s sovereign default risks, while higher capital requirements and the backing of the solvency support facility would restore banks’ creditworthiness.

  • 4. The policy program for the member country provides a reasonably strong prospect of success, including not only the member’s adjustment plans but also its institutional and political capacity to deliver that adjustment. Staff considered the adjustment efforts in the program to be significant and the institutional capacity adequate. Since Portugal was facing elections during program negotiations, there were obvious concerns about the durability of a program under a new government. These were allayed by getting commitments from the main opposition parties to support the program’s objectives and key measures.

Key Events and Market Sentiment

  • Following approval of the EFF, in May 2011, sentiment continued to worsen through June elections and ratings downgrades. Spreads declined only around the time of the second bailout of Greece in July, when the EFSF extended maturities and cut rates on financing, including for Portugal.

  • The successful completion of the 1st review of the EFF and the release of the 2012 budget, which deepened civil servant wage and pension cuts, failed to improve sentiment. However, the EU summit in October 2011, which strengthened firewalls for financial stability in the region, and the first ECB LTRO in February 2012 provided some temporary respite to spreads.

  • Bond yields reached euro-era highs at end-January 2012, as many investors priced in a default amid fears that debt holders could suffer from a restructuring deal with Greece.

  • The second ECB LTRO set the stage for improvement in sentiment which continued thorough completion of the 4th EFF review until the ECB announcement (July) and the OMT (August) cemented the decisive downward trend in spreads.

  • Volatility increased in September 2012 around the government’s announcement of fiscal devaluation measures, which were abandoned in the wake of large protests.

  • The first Constitutional Court ruling (July 2012) and the issuance of the first 10-year bond (May 2013) during the program took place in the context of declining spreads. Sentiment worsened again at the time of the resignation of the Minister of Finance and the 7th review, but spreads then remained reasonably stable in the face of successive Court rulings against program measures and a delay in concluding the 8th review.

  • Portugal successfully floated a 5-year bond in January 2013, followed by a 10-year bond in May, as growth began to recover and despite evident reform fatigue; in January 2014, a further 5-year bond issue was successful, and in April a 10-year issue, before Portugal exited the program.

uA01fig02

Portugal: 7-Year Government Bond Yield, 2011-14

(Percent)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Program Implementation and Outturn

Portugal’s program succeeded well as a response to a capital account crisis. Flow imbalances were decisively reduced and Portugal regained market access early. However, at the end of the program, public and private debt exceeded program targets, the fragility of banks had become more evident, and growth continued to disappoint. Already at end-program, important measures had begun to be reversed. Hence—as an assessment—the program worked as a crisis deterrent and holding operation, but did not make inroads to entrench fiscal and financial sustainability.

A. Macroeconomic Outcomes and Political Developments

19. The program impressively reduced flow imbalances, but did little to lower debt stocks. The current account deficit closed more quickly than programmed, shrinking by more than 10 percent of GDP to reach a small surplus by 2013. The structural fiscal deficit improved by 7 percentage points. But debt levels remained problematically high, reflecting financing needs during the program, unexpected additions to public debt, a deeper-than-anticipated downturn, and a weaker recovery. Private debt (unconsolidated) declined modestly from 247 percent of GDP in 2010 to 239 percent in 2014, while public debt rose from 96 percent of GDP in 2010 to 130 percent in 2014.

20. The downturn during 2012-13 proved larger than expected. The program had projected a contraction of close to 2 percent in 2012, followed by a growth rebound from 2013-16 led by recovery in investment. In fact, output shrank more in 2012 and continued to fall in 2013, requiring a revision to program growth projections in the 2nd and 5th reviews.7 The contraction was followed by a tepid recovery, leaving real GDP in 2014 6 percent lower than in 2010, compared to a decrease of 0.5 percent projected in the program request. The sharper downturn was accompanied by greater unemployment, nearly 2 percentage points higher than expected at end-2014 (Figure 3).

Figure 3.
Figure 3.

Macro Developments—Program and Outturn

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: Portugal EFF program reviews

21. The deeper downturn reflected a much weaker external environment and a more pronounced impact of deleveraging than envisaged. The larger than expected downturn was not specific to Portugal, as the Fund’s euro area forecast turned out to be unduly optimistic over the crisis period. At the program request, the euro area had been projected to grow by nearly 2 percent annually from 2012-16, but it contracted by nearly 1.5 percent during 2012-13, and remained sluggish in 2014. The recession was particularly severe in Spain, by far Portugal’s largest trading partner. The decline in credit growth as banks deleveraged also proved larger than expected, contributing to a sharp fall in private investment that continued into 2013. This was compounded by the ongoing fiscal adjustment.

22. The sharper drop in investment contributed to a much faster adjustment in the current account than expected. At program request the adjustment in the current account was projected at about 3-4 percent of GDP, but the deficit shrank by almost 10 percent of GDP by 2014 (bottom left graph in Figure 3). The adjustment had been expected to occur through import compression due to shrinking demand, with a later recovery export-led. But in sharp contrast to other internal devaluations (Greece, Spain, Ireland) two-thirds took place through export growth (Figure 4).8 Export growth benefitted from a redirection of trade flows to new markets, a shift by domestic producers to export markets, and expanding tourism (IMF, 2013). However, a recent study shows that the large export recovery was linked also to re-exports of oil and other imported intermediate inputs, at least up to 2013 (Gershenson et al., 2016) raising questions about the sustainability of the recovery.

Figure 4.
Figure 4.

External Adjustment

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

23. The authorities demonstrated strong ownership of the program, particularly in the first 18 months. In the face of setbacks to policies and weaker growth, they succeeded in meeting all fiscal targets, with steady progress on implementation of structural reforms. Ownership appeared to weaken as financing pressures eased, however, with the ECB’s announcement of a new outright monetary transactions program that included the prospect of sovereign debt purchases helping to facilitate Portugal’s return to international capital markets in 2013. The political consensus in support of the program became increasingly fractured following widespread public resistance to the fiscal devaluation proposed by the government in September 2012, with mounting opposition from vested interests to further reforms. Program implementation also faced significant obstacles from a series of Constitutional Court rulings, which overruled key measures in the budgets for 2013 and 2014. The authorities were able to identify alternative measures in response to the court rulings to meet the fiscal target for 2013, but with financing conditions increasingly favorable, opted to let the arrangement expire without finding additional measures needed to complete the final review.

B. Fiscal Policy

24. The program achieved significant fiscal consolidation, although it fell well short of target. The fiscal balance improved by about 6½ percent of GDP (excluding pension fund transfers—see below); however, this adjustment fell short of returning Portugal to compliance with the SGP’s 3 percent target. The composition of consolidation was less supportive of growth than planned, with efforts shifting to the revenue side as spending cuts failed. It was also less front-loaded than designed, with adjustment postponed in the face of policy slippages. The debt to GDP ratio stabilized only at the end of the program, and at a much higher level than originally envisaged.

uA01fig03

Change in Structural Balance

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: Portugal EFF program reviews.

25. The program softened fiscal targets as the growth downturn exceeded expectations. The program accommodated a smaller adjustment for 2011 that resulted from policy slippages, although it planned to compensate in 2012. However, later, the targets for 2012 and 2013 were relaxed as the recession was worse than envisaged (Figure 5). The budget target for 2014 was 4 percent of GDP, but without substitute measures to offset the impact of the Constitutional Court’s reversal of measures on pensions and unemployment subsidies, the deficit widened to 7.2 percent of GDP after the program (including one-off support for banks).

Figure 5.
Figure 5.

Fiscal Developments—Program and Outturn

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

26. Fiscal slippages and new replacement measures were a recurrent feature of the program from the start. In 2011 the slippages came mainly from expenditure, and while some were related to one-off support for SOEs and PPPs, others were more permanent. They were compensated mainly by a transfer of bank pension assets (Box 2), meaning that adjustment was postponed to the second year. In 2012, the slippages shifted to the revenue side as growth disappointed and imports dropped more than expected, hurting VAT revenue. In 2013, reversal of spending measures required additional revenue effort.

  • Revenue measures were of variable quality. The measures that allowed the government to meet the 2011 fiscal target were temporary, including a surcharge on 2011 personal income taxes, and the pension asset transfer. The tax measures taken in 2012 were growth-friendly and good quality, with efforts focused on broadening bases, including by increasing the VAT base subject to the standard rate and reducing PIT and CIT tax expenditures—but they failed to generate as much revenue as envisaged. PIT rate increases in 2013 were less growth-friendly but permanent and more effective in raising revenue.

  • On the spending side, replacement measures concentrated on the wage bill and pensions but were reversed by the constitutional court. Overruns in the first year, on goods and services outlays and support to SOEs and PPPs, caused the government to seek larger cuts in the civil service wage bill—including a two-year suspension of 13th and 14th monthly wages and pensions. The cuts remained in place during 2012, but were reversed by the constitutional court for 2013. Moreover, despite measures to improve targeting of social transfers and efforts on health sector and SOE reforms, non-wage spending increased in 2013. The 2014 budget envisaged cuts to pension and lower unemployment and sickness benefits, but these were also overturned by the Constitutional Court.

27. The planned fiscal devaluation was not feasible. The program envisaged reducing reliance on labor taxes—offsetting revenue losses through higher consumption taxes and expenditure cuts—to promote an internal devaluation. Given spending slippages and weak VAT performance, the authorities’ proposal changed to shifting social security contributions from employers to employees. In the face of widespread protests, this was quickly discarded.

The Budgetary Impact of One-Off Measures

Underlying fiscal adjustment in Portugal moved differently to the headline numbers, because some program targets were met by large temporary revenue and expenditure measures. By far the largest item was the temporary revenue boost from the transfer of pension fund assets to the general government, but others included the temporary PIT surcharge on 2011 taxable income, and the issue of 4G licenses. On the expenditure side, one-offs included the recapitalization cost for BPN, CGD, and Banif, as well as a capital operation for Madeira. Some operations—like the recapitalization—reflected risks that materialized; others—like the PIT surcharge—were pursued to compensate for such temporary slippages. The use of one-offs to meet the targets contributed to explaining why, at the end of the program, Portugal had adjusted (in structural balance terms) by 7 percent of GDP rather than the 9¾ percent of GDP envisaged in the program.

uA01fig04

Budget Balance One-Offs

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: IMF staff calculations.

The pension fund transfer operation was largely an accounting stratagem to meet targets. The government transferred all assets and liabilities related to retired bank employees into the fiscal accounts. The transfer of assets entailed a reduction in the deficit according to both ESA95 and GFSM 2001 accounting rules (because the future liabilities offsetting the current asset inflow were not classified as deficit-creating). Nonetheless, both staff and the authorities recognized that this was not an appropriate fiscal measure as it would not improve the underlying fiscal position. Less apparent at the time was the fact that, by delaying adjustment to 2012, the fiscal impulse would be more negative then, than planned in the program.

The use of pension fund transfers to meet deficit targets was excluded from 2012 onwards. Portugal Telecom pension assets of 1.6 percent of GDP had previously been transferred in 2010, and the program request envisaged winding down such transfers to only 0.2 percent of GDP in 2011. During the second review, the authorities informed the Fund that they intended to transfer 1.9 percent of GDP. The Fund accepted the measure, but amended the definition on the floor on the general government cash balance to exclude revenues from reclassification of pension funds from 2012 on. The final transfer for 2011 was much larger—at 3.5 percent of GDP—which allowed the targets to be met despite significant slippages. The consequence was a more negative fiscal impulse for 2012 than in the original program.

C. Debt and Financing

28. Debt stabilized only at the end of the program. The program envisaged that public debt would peak at 115 percent of GDP in 2013 (rising from 93 percent in 2010), declining to 113 percent by end-2014—with staff candid about the many risks associated with this plan. In the event, debt peaked at 130 percent of GDP in 2014, with sustainability predicated on future higher primary surpluses.

29. The higher debt reflected weaker fiscal outturns and lower growth, but also additional bank recapitalization costs. Portugal raised EUR 31 billion during 2011-13 in additional net placements (41 billion from residents). This was needed to pay for the weaker fiscal outturns, for one-offs such as the recapitalization of CGD (outside the BSSF) and recognition of debt previously outside the perimeter of general government, and to accommodate the lower than expected roll-over by non-residents. The persistence of shocks, including further reversals of measures in 2014, and uncertainty about future bank needs meant that, at the end of the program, the debt outlook remained fragile.

uA01fig05

General Government Debt

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: Portugal EFF program reviews.

D. Financial Sector

30. The broad goals of the program were achieved, but the banking system emerged in a weak financial position. Scarce capital, low profitability, and growing impairments as the nontradables sector retrenched, left banks fragile. Banks met the higher capital targets in the program,9 including through raising some new equity from private investors or tapping into BSSF support. Banks also deleveraged significantly, though at a moderate rate supported by continued significant reliance on ECB financing (Figure 6). In 2013, several banks (including BES) returned to debt markets. However, limits to capital increases became apparent, as bank profitability stagnated (Figure 7). Credit contracted dramatically and, as the economic environment deteriorated, NPLs rose sharply. The third largest bank, BES, failed shortly after the program and Banif, a Madeira-based bank, was intervened in late 2015 (after earlier recommendations for what staff believed would be a less costly resolution option).

Figure 6.
Figure 6.

The Liquidity Squeeze and the ECB to the Rescue

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Figure 7.
Figure 7.

The Challenging Quest to Raise Capital

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

31. Bank deleveraging affected small enterprises most. Rationed credit went to more creditworthy customers (large corporates) while more vulnerable (and less credit-worthy) players got less financing. About EUR 17 billion of the total EUR 25 billion contraction in net credit over the program came from SMEs and micro-enterprises (many of which were highly indebted or inefficient), which led to a steep increase in delinquencies (Figure 8). Fearing a credit crunch in an important sector of the economy (these enterprises contributed nearly three-quarters of value-added and employment), the authorities introduced support measures, but they helped little to mitigate the lending decline.

Figure 8.
Figure 8.

The Bank-Corporate Negative Feedback Loop

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

32. The new infrastructure for facilitating corporate debt restructuring yielded modest results. The in-court and out-of-court corporate debt restructuring processes put in place under the program were in line with best international practice. But the limited capacity of banks to absorb losses lowered incentives to engage in loan restructuring. In early-2014, only about 1,800 companies were undergoing restructuring processes, out of over 70,000 vulnerable companies (EBITDA lower than interest expense). Other impediments to faster restructuring included lack of creditor coordination and new financing, limited judicial capacity, and the preference of both banks and companies to avoid taking losses, in the hope that growth would eventually ‘lift all boats’.10 Near the end of the program, staff called for an overhaul of the corporate debt restructuring strategy.

E. Structural Reforms and Competitiveness

33. The large structural reform program was formally well-implemented, but fell short of closing Portugal’s competitiveness gap. All structural benchmarks were met, except for fiscal devaluation. Unit labor costs came down—especially in the tradable goods sector but also in construction and distribution—and important steps were taken to flexibilize the labor market. However, high unemployment persisted (Figure 9), partly because product market reforms remained incomplete and final prices declined little.

Figure 9.
Figure 9.

Labor Market Conditions

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

34. Early implementation of labor market reforms was impressive but, with delays and legal obstacles, ownership weakened. In the public sector, wages were reduced at first, but later reconstituted following Constitutional Court rejection. And between July 2012 and September 2013, Labor Code revisions affecting working time and employment protection were reversed. Some active labor market policies were implemented successfully: training programs for youth and the unemployed were shortened and better targeted. Portugal managed to create jobs and to reduce unemployment, which however remained high (Figure 9). However, deeper reforms to lower unemployment benefits and the cost of dismissals, and address the generosity of severance payments, continue to face political resistance, job protection remains among the highest in the EU, and the new jobs are disproportionately in temporary contracts.

35. Product market reforms proved difficult. Reforms of regulations required extensive knowledge of the market and legal framework, took time, were implemented only partly and, perhaps in any case, would not have generated immediate benefits in a downturn. Vested interests prevailed (a risk highlighted in the program document). Energy sector reforms encountered most obstacles. The goal was to eliminate excessive rents, particularly in electricity. But renegotiation of contracts proved difficult, partly because it would have either affected future privatization revenues (which the authorities preferred to maximize) or penalized new shareholders. A levy on energy operators was later introduced to claw back part of the rents.

36. Judicial reforms were substantive, though still considered work-in-progress. The crisis provided the right context for moving forward reforms already contemplated by the authorities earlier. Nonetheless, overhauling the legal system (the new Code of Civil Procedure and the Judicial Organization Act) was a protracted process. The eventual reforms reduced administrative burden, and improved efficiency by increasing the specialization of judges and introducing performance-based models. Court organization was streamlined and modernized to reflect international best practice. The authorities resolved a large backlog of enforcement court cases and, by increasing processing speed, prevented the accumulation of new ones. A new system of garnishing bank accounts recovered 0.2 percent of GDP in payments from old cases within 15 months of application (Pompe and Bergthaler, 2015). Elimination of the requirement of a judicial decision for VAT recovery unburdened the courts. And a new supervisory framework and incentive-based fee structure for enforcement agents improved the efficiency and effectiveness of debt enforcement.

Structural Conditionality—Design and Implementation

Portugal’s structural conditionality was not an outlier compared with other European programs relying on internal devaluation, but it was more frontloaded and the reform agenda was extremely broad. Conditionality was more heavily weighted towards fiscal reforms, but the strategy in the MEFP (and more so in the MOU to European partners) covered a far broader spectrum of reform commitments. Conditionality was more frontloaded compared with other programs, with about half of all prior actions and benchmarks to be implemented in the first year. The 22 conditions set at program request were augmented to 55 by end-program. Benchmarks do not appear to have set excessively detailed guidance, suggesting perceived strong program ownership. See Table 3 for a full list of prior actions and benchmarks.

uA01fig06

Cumulative Number of Conditions per Review

(Structural benchmarks and prior actions)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: Portugal EFF program reviews.
uA01fig07

Share of Front-Loaded Measure in Total

(Percent)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: Portugal EFF program reviews.

The competitiveness agenda was extensive, due to the need to deliver a decline in economy-wide costs. Some main objectives were lowering government interference in business, upgrading the regulatory framework, divesting SOEs, fostering competition, and reforming civil and commercial claims enforcement. A key focus was to improve labor market flexibility, productivity and skill mix. Benchmarks included measures to reduce severance payments and amend the insolvency law, eliminate golden shares, and revise the competition law. Some important reforms were mainly specified in the MEFP (rather than in conditionality), notably in product markets to open up professional services and reduce rents in energy, telecommunications and transportation.

On the fiscal side, the focus was on addressing fiscal risk from public enterprises and PPPs and improving public financial management and transparency. At the outset, the emphasis of conditionality was on stock-taking and institution building. Benchmarks on assessing the health of SOEs, stock-taking of arrears, and evaluating PPP contracts would provide information needed to implement reforms in commitment control systems, elaborate a multi-year budget strategy, and revise the regional finance law etc. Conditionality evolved over the program to incorporate knowledge from technical assistance.

Financial conditionality was used to implement the agenda to strengthen bank balance sheets and the legal framework for oversight, the safety net, and intervention. A key prior action was to increase banks’ capital requirements.

Compliance with the benchmarks was more than satisfactory, but implementation encountered obstacles. Of the total 55 measures, over 80 percent were fully met and on time. A few reforms were postponed to later reviews, when the original timetable was later judged unduly optimistic. Relatively more frustrating were delays in SOE reform and arrears clearance, where measures proved difficult, often requiring passage of laws and intermediate steps to effectively implement the changes.

Reform outcomes were mixed, partly reflecting the adverse environment. Domestic arrears continued to accumulate during the program, mostly at the subnational level, where implementation of measures on commitment control and the regional finance law faced hurdles. Moreover, significant price rigidities from existing contracts, the renegotiation of which was difficult and lengthy, prevented a reduction in production costs in product markets (in electricity and transport). Declining demand for services reflecting the severe recession may also have impeded companies from passing on falling costs to consumers.

37. There was some improvement in cost-competitiveness, but mainly on the back of declining wages and employment. The ULC-REER is estimated to have declined by 7 percent during the program, narrowing the wage competitiveness gap with the euro area. ULC gains were partly triggered by wage restraint in the public sector; decentralization of wage bargaining did not lead to negotiation of lower wage contracts.11 But labor-shedding was the main driver of ULC gains; unemployment remained high.

38. Because of nominal rigidities in product markets, consumer prices fell less than ULCs, reducing competitiveness gains. To some extent, the gains from labor market measures were undermined. Product market reforms were designed to reduce production costs, particularly for exporters. But upward pressures on electricity prices from the structure of purchasing contracts with generators did not deliver the needed decline in input prices. The benefits of reforms in the ports sector were also not fully passed through to end-users, as they too required a renegotiation of contracts with existing concessionaires. As the program progressed, staff expressed concern about rigidities, persistent excessive rents in non-tradable sectors, and continued high labor protection (despite much progress), signaling the need to implement additional measures before the reforms could trickle down to final prices.

uA01fig08

Harmonized Competitiveness Indicators

(Index: 2010=100)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: European Central Bank.

Program Design Issues

Portugal’s program was an attempt to combine crisis response with fundamental structural change, while respecting the constraints of currency union membership and minimizing the cost to growth. At the end of the program, tepid growth had returned, but public and private debt remained high and the financial sector remained shaky. Looking forward, these fragilities are more marked and sustainability still in question. Did the immense balancing act put unreasonable demands on program design?

A. Could Sustainability Have Been Achieved by Better Program Design?

39. With the benefit of hindsight, there were a number of reasons why the program fell short of achieving all its goals. These are listed here and examined in more detail below:

  • The severity of the recession was underestimated.

  • Actual fiscal adjustment was less than first targeted in the program, partly because targets were amended to support activity, partly because measures were temporary or later reversed, and partly because the composition of compromise measures was not ideal.

  • Internal devaluation is a long and difficult process, by its nature requiring changes in relative incomes, likely to entail higher unemployment while factors shift from sheltered to competitive sectors, and with balance sheet effects analogous to when the exchange rate is devalued. When rigidities are entrenched and private sector balance sheets are seriously weakened, restoration of competitiveness could take far longer than a three-year EFF.

  • Political ownership of the program eroded, partly due to fatigue from the string of new measures needed to compensate for slippages or reversals, and partly because financial incentives to complete the adjustment weakened with Portugal’s return to markets and expectation of an easier financing environment in the context of the euro area’s assurance of ECB support ‘whatever it takes’. Frontloading of the program was appropriate but also had the downside of leaving fewer incentives in place to finish the reforms envisaged under the program.

  • The choice not to seek sovereign debt restructuring. This strengthened Portugal’s credibility and facilitated early re-access to markets, but left it with a higher debt overhang than at the start of the program.

  • The need to deleverage multiple sectors of the economy simultaneously. The constraints on credit growth of beginning to unwind such high leverage were not well-understood. But even if they had been, it is not clear that a more aggressive bank and corporate restructuring process would have been feasible (given the lack of effective instruments) or desirable (given that the disruption to growth would have been more abrupt and since capacity to fight crisis on an additional front could not be assured.)

B. Should the Larger Economic Downturn Have Been Expected?

40. The growth shortfall relative to projections created difficulties for the program. The program envisaged a shallow recession in 2011-12 and return to growth in 2013. In the event, 2011 was slightly better than predicted, but the recession in 2012 was significantly more severe (-4 percent growth rather than -1.8 percent) and positive annual growth did not return until 2014.12 The shortfall in growth was costly to the program, in terms of lower yields from fiscal measures, loss of credibility, and social fatigue in the face of replacement-measure announcements. The ambition of the fiscal program had to be repeatedly scaled back.

Portugal: Macro Indicators - Expectations and Reality

article image

Fiscal outturn as reported under ESA2010. Under ESA2010, the transfer of pension funds from the private to the public sector in 2010 and 2011 is recorded as a financial transaction rather than a revenue item, increasing the headline fiscal deficit. The 2014 outturn also includes the cost of banking sector support after the program amounting to 2.8 percent of GDP.

41. There were a number of reasons for the worse outturn—some but not all of which could have been taken into account.

  • The effective delay in fiscal contraction to 2012, since the use of the pension transfer to meet the 2011 target meant a smaller withdrawal of demand than envisaged in the program. This helped support growth in 2011 but implied that the tightening in 2012 was more than the headline numbers suggest. Growth projections were amended in the second review to reflect this.

  • The intensification of bank deleveraging forced by the sudden stop in external financing. Private credit was much tighter than programmed, private savings much higher, and the current account deficit closed far more rapidly than programmed. This worsened the blockage to growth created by an over-leveraged private sector (that should but may not have been reflected in the program baseline). Arguably staff could have recognized the tension between worsening conditions in the private sector and maintaining credit, but the analysis of macro-financial feedback loops was still a new issue for economists.

  • The lower-than-expected activity in the euro area from 2012 on.

  • The depressing impact of internal devaluation on growth. As discussed below, in economies where prices are sticky downwards, it is unlikely that internal devaluation could be achieved without lowering growth. The timing of the impact would be uncertain, and other program features aimed at cushioning it, but it could be argued that the programmed growth path did not fully reflect the implications of the mechanism for delivering competitiveness gains.

It must be recognized that these factors were difficult to foresee, and that estimates of their impact were exceptionally uncertain. At the beginning of the program, Fund projections were below consensus—but with forecast errors no worse in Portugal than in the rest of the euro area. 13 Later, however, growth could have been marked down further, to reflect the general belief in Portugal and among other Troika members that the projections were over-optimistic, and to respond to slippages.

C. Were the Size and Pace of Fiscal Adjustment Appropriate?

42. The lower-than-projected growth raises the question of whether fiscal adjustment should have been less or more gradual. Findings (WEO 2012) that fiscal multipliers for euro area countries were higher than assumed in the Portuguese program unleashed concern in Portugal that the fiscal program was too ambitious.14

43. A large front-loaded fiscal adjustment was a core element of the Troika strategy. Several compelling considerations constrained the size and pace of adjustment:

  • Both the authorities and European partners wanted to establish the credibility of the program (and differentiate Portugal’s situation from Greece’s) by returning the deficit to compliance with the 3 percent of GDP SGP target as rapidly as possible.

  • Slower adjustment (and higher financing) would have raised debt further to levels even less credibly stable.

  • Disbursements under the program were frontloaded to ensure Portugal could meet high near-term financing needs. Adjustment needed to be comparably frontloaded to avoid moral hazard (loss of commitment to adjustment after the money was disbursed).

The compromise fiscal path negotiated by Fund staff (with the deficit reaching 3 percent of GDP a year later than in the Stability Program) attempted to strike the right balance between respecting these concerns and the desire to support growth. Moreover, in the face of the deeper-than-envisaged recession, fiscal targets were relaxed. Further relaxation of the fiscal targets without imperiling debt sustainability would have required revisiting the decision not to restructure debt. As discussed below, this was unanimously ruled out by the authorities, the Fund and European partners.

The Debate about Fiscal Multipliers

For Portugal, the recession was certainly deeper and more protracted than envisaged at the time of the program request. Several studies have stressed that short-term multipliers depend on the state of the economy (Tagkalakis 2008, Auerbach and Gorodnichenko 2010, Corsetti et al 2012, Baum et al 2012), with larger multipliers during recessions. For Portugal, a calibrated DSGE model (Castro et al 2013) also suggests larger multipliers during recessions, particularly for spending-based consolidations.

Blanchard and Leigh (2013) found that, for European economies, growth forecast errors were negatively correlated with the forecast fiscal consolidation, and argued that this reflected forecasters’ underestimation of fiscal multipliers. Their results suggested that one-year multipliers were one unit higher in 2010-11; on the assumption that forecasters used ½, this suggested that the actual multiplier was 1½.

While the results for 2010-11 grabbed headlines in the global debate about fiscal austerity, results for the subsequent years—during Portugal’s crisis—were in the ⅓-½ range and had weaker statistical significance. Möhlmann and Suyker (2015) found even weaker significance using new data vintages to calculate the forecast errors.

Further, more recent research suggests the error in the multiplier might have been smaller:

Lall et al (2016, forthcoming) argue that such an approach does not control for shocks to aggregate supply, and hence overstates the error in fiscal multipliers. To capture supply shocks, they used forecast errors for the output gap rather than for growth and found smaller errors in the multiplier for 2010-11 (by about 40-45 percent). Their estimates of multipliers of ½ would imply that the implied multipliers used by forecasters were close to zero. Huidrom et al (2016) argue that multipliers depend on fiscal positions, and found significantly lower values for countries with high debt levels.

The lack of consensus in this debate is not a surprise. The two-way causality between output and fiscal policy—due to automatic stabilizers, countercyclical fiscal policy—is a significant challenge when trying to measure fiscal multipliers. While researchers have tried to identify changes in fiscal policy not induced by macro-conditions or by extracting exogenous components from observed fiscal outcomes, there is no broad consensus on how to do it.

44. As regards the size of fiscal adjustment, less adjustment would obviously have supported demand but the jury is still out on how big the impact would have been on growth. The fact that, for a given level of fiscal adjustment, measured ex-post multipliers showed weaker activity than projected at the time of the program does not mean that fiscal adjustment ‘caused’ the growth decline. The other factors discussed above also played key roles in depressing demand. The substitution of public demand for depressed private demand would have supported activity, but not necessarily with symmetrically large multipliers, given the other constraining factors and the signaling impact on confidence. As discussed in Box 4, fiscal impact would need to be assessed controlling for all other effects on growth; recent research efforts to do so suggest smaller (though still positive) multipliers.

Trade-Offs in the Pace of Fiscal Adjustment

The pace of fiscal adjustment influences the depth of recession and level of debt. An important question for Portugal was whether delaying consolidation—and the associated growth downturn—until later would have left the economy better off. Portugal’s debt and tax to GDP ratios implied that, for sufficiently high multipliers, a more gradual consolidation could reduce debt ratios in the short run. But, regardless of the multiplier, back-loading consolidation would leave Portugal with higher debt levels in the medium term.

This point is illustrated by the following simulation assuming low (0.5) and high (1.5) short term multipliers, and shifting the fiscal consolidation by ½ percent from year 3 to year 4. For a high enough multiplier, the higher output from delayed consolidation could lower the short-term debt to GDP ratio. But the eventual consolidation will quickly dissipate the short-term effect on output; the medium-term debt ratio will be higher—since the delay had to be financed—and roughly the same regardless of the multiplier

uA01fig09

Output

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: Portugal EFF program reviews.
uA01fig10

Real GDP

(Year-on-year percent change)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: Portugal EFF program reviews.
uA01fig11

Debt

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Source: Portugal EFF program reviews.

A different perspective is to focus on the fact that short-term multipliers vary over time, and it makes sense to adjust when they are lower—to the extent that financing is available. Other things equal, back-loading makes sense if multipliers are expected to decline during the program. This was not discussed explicitly in Portugal, as the justification for a front-loaded adjustment focused on addressing the debt overhang and rebuilding credibility. That said, at the start of the program, several factors suggested that multipliers could increase: the output gap was not large and was envisaged to widen during the program; and the ECB was tightening (although that was later reversed).

The trade-off will also depend on the longer-term impact of higher debt on growth, though this is not reflected in the simulation. The literature on the long-term effects of fiscal policy is still incipient. Higher debt has been linked to slower growth and investment (Fiscal Monitor April 2013), but on the other hand, De Long and Summers (2012) argue that there are ratchet effects from reducing physical and human capital investment, and therefore long-term gains from avoiding a deeper recession.

45. As regards the pace of adjustment, even-loading (as opposed to front- or back-loading) could have generated a slightly shallower recession. Simulations suggest that spreading the fiscal adjustment evenly could have cushioned the decline in growth rates in the early years and achieved a shallower recession, but even with high multipliers the gain would have been small, and would have come at the cost of higher debt. Back loading would have alleviated the recession little or not at all, and would have required more—and more front-loaded—support from the Troika, which would have increased moral hazard. Conceivably Portugal could have earned some credibility not tied to immediate deficit cuts by enacting impressive fiscal structural reform. Civil service reform was one such measure prioritized in the program—but the political support needed to pursue it was never feasible.

46. The composition of the fiscal package, as implemented, could have been more supportive of growth and competitiveness. The expenditure-rationalizing measures in the original fiscal strategy aimed not only to achieve adjustment but, by reducing the premium in public sector compensation, eliminate disincentives to wage rebalancing in the rest of the economy and help lower costs. Moreover, expenditure-led adjustments have robustly been found to be more successful and more supportive of growth than revenue-side packages. Likewise, a socially-tolerable fiscal devaluation could have eased financial pressures on corporations by a possible 1 percent of GDP at a time when these pressures were most inimical to growth. The Constitutional Court’s reversal of several important cost-cutting measures, and their replacement by compromise tax measures, reduced the quality of the package and exacerbated uncertainty about the future tax climate. The unbudgeted financing needs of SOEs and local governments crowded out resources from the private sector (since, beyond an allowance for contingencies in the original program, these were financed through the banking system).

D. Should the Debt Have Been Restructured?

47. High debt limited the scope for more gradual fiscal adjustment, but sovereign debt restructuring was ruled out from the beginning. There was unanimity across the Troika and the authorities not to pursue a restructuring of privately-held debt. The government put high value on maintaining Portugal’s credit reputation, with a view to regaining market access with minimum delay; there were also concerns that a restructuring would have generated substantial turmoil in European financial systems at a time of great uncertainty, including creating adverse spillbacks to Portugal. More generally, differentiating Portugal’s (and Ireland’s) problems from Greece’s was seen as important for reducing the tail risk of a breakup of the euro area. Despite the serious risks acknowledged in the program, the high credibility that came from the government’s strong ownership at the outset gave confidence that Portugal’s financial difficulties could be resolved without restructuring.

48. Staff reopened—within the Fund—a discussion of private sector involvement (PSI) in early 2012. Greece had restructured, Portugal had been downgraded to junk status by the major rating agencies, and spreads were at an all-time high (see Box 1). Moreover, the timing was relatively propitious, because European firewalls had been strengthened but large private debt redemptions were still in the future (so savings from PSI could have been substantive). 15 However, the program was on track, and the other considerations above were still pertinent, so PSI was not seen to be in Portugal’s best interest. Spreads were never as high again; the EU extended the maturities on its share of Portugal’s program debt in April 2013; large redemptions in the second half of 2013 significantly reduced PSI-able debt; and Portugal regained international market access in early 2013.

49. This report shares the view of the Portuguese authorities and the Troika that debt restructuring at program outset would have created more damage than it was worth. Given the large amounts of debt held by domestic and euro area banks, such an operation would likely have been disruptive, both for domestic and regional stability. Moreover, when a later opportunity arose, the benefits did not seem to outweigh the costs, in a context where re-establishment of market access was in sight. As discussed in Box 6, the concrete short-to-medium term costs of losing market access and of higher risk premia dominated the uncertain gains from even a relatively substantial debt restructuring (still more so from an EU perspective if these risk premia spilled over to other euro area countries). Lower (long-term) growth than projected in the program would have increased expected gains from the operation, but these gains could easily be swamped by the higher risk premia likely to be associated with such a poor growth outlook.

50. A shallower reprofiling at the beginning of the program might have been less costly but would also have brought fewer benefits. Locking in PSI-able debt by extending maturities but without face-value debt reduction) might not have impeded market access or worsened terms for so long a period, but there would have been no significant improvement in debt sustainability; 16 moreover, as a credit event and in the absence of firewalls, it could have triggered comparable spillovers to a restructuring with face-value haircut.

Sovereign Debt Restructuring—No Easy Choices

The paper concludes that sovereign debt restructuring—meaning a stock operation to meaningfully reduce the size of sovereign debt—was never a realistic option in the program. This box explains this conclusion, laying out the choices facing Portugal, using illustrative but realistic numbers.

uA01fig12

Estimates of PSI-able debt

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Sources: Authorities and staff estimates.

The decision to restructure sovereign debt is a cost-benefit analysis. The benefit is the debt service relief, and the impact of the haircut on the debt-stabilizing primary balance and on growth. The cost is the loss of market access, the higher risk premium (for both sovereign and private debt) after market access is regained—and, in Portugal’s case, externalities and spillbacks if the restructuring had adverse spillovers to the rest of the euro area.

Portugal’s potential gross gains from a debt write-down could have been significant, but would probably have been offset by high costs. Estimating the size of benefits is complex, as serious uncertainty surrounds the outcome of a debt operation. To illustrate the possible size of debt relief, Edwards’ (2015) results can be used to forecast a plausible range of haircut and consequent range for debt reduction.1 On an estimate of PSI-able debt,2 the implied debt reduction would have been 18 percent of GDP (average of a range of 9-27 percent) at the beginning of the program, declining to 12 percent of GDP (6-18 percent) if postponed until the end (after private creditors exited)—meaning that Portugal would have been in the upper end of the distribution of countries’ saving from debt restructuring.

However, costs were also likely to have been large. The most concrete costs included:

  • (i) a delay in the return to market access (forgone financing, net of the debt service relief made available by the restructuring). Cruces and Trebesch (2013) find that it took countries on average 5 years to re-access markets after final restructuring events, and longer after higher haircuts. After reaccessing the market in 2013, Portugal more than covered its total gross financing needs (the issues were oversubscribed)—allowing fiscal targets to be loosened without increasing official financing; and

  • (ii) higher sovereign spreads, at least in the short-run, particularly for large haircuts. Portugal’s high foreign debt—exceeding 200 percent of GDP (private as well as public)—implies that even a small prolonged increase in sovereign risk premium would have been costly. Estimates from Cruces and Trebesch, that a 40 percent haircut could increase sovereign spreads for 6-7 years by about 55 bps, suggest that interest on Portugal’s external debt could be about 6¾ percent of GDP higher over the 7 years following a restructuring.3

uA01fig13

Estimates of plausible ranges for haircuts and savings

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Sources: Authorities and staff estimates based on Edwards(2015)Note: Range built using the point estimate and +/− one standard error.
uA01fig14

Histogram of PSI debt restructuring savings

(In percent of GDP)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Sources: Trebesch and Cruces (2013), WEO and staff estimates.

The authorities were also concerned about the impact of restructuring on GDP and FDI. Das (2012) found an association with declines in GDP (2-5 percent loss) and FDI (2 percent of GDP loss), but establishing causation is not simple; moreover deferring restructuring would not avoid these costs if debt were truly unsustainable.

A fundamental question is how much restructuring would have strengthened debt sustainability. Given the low interest rate environment, the plausible impact for Portugal seems modest. The gain (i.e., the decline in the needed primary balance) depends on the long-term impact on the risk-premium (which is quite uncertain) and the debt dynamics generated by the difference in interest and growth (r minus g). In the program baseline, long-term growth was projected to be 2 percent and the interest rate 3.2 percent. In this case, a debt reduction of 18 percent of GDP would reduce the debt-stabilizing primary balance by around 0.25 percent of GDP (or by more if starting interest rates were higher or growth were lower). However, if the debt reduction raised the permanent risk-premium by 20 basis points, all gains would be wiped out (40bp if starting rates were 1 percent higher or long-term growth were 1 percent lower).

uA01fig15

Change in the debt-stabilizing primary balance

(In percent of GDP)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Sources: Staff estimates.

These illustrative magnitudes give insight into Portugal’s decisions not to restructure sovereign debt. Looking first at short-run liquidity concerns, the loss in market access and higher public and private interest bill loomed large relative to the potential debt relief. The scale of debt reduction achievable is uncertain, but comparisons with other cases suggest that it would likely have yielded at best a modest improvement in debt sustainability. Added to these comparisons is the fact that, at the outset of the program, a major concern from the perspective of European partners was the fear of spillovers to other euro area countries (i.e., a multiplication of higher spreads and possible broader loss of market access). Given these baseline magnitudes and the uncertainty surrounding them, there was never a time when embarking on debt restructuring would have been clearly beneficial (though, as usual, the beginning of the program would have been the most advantageous time).

Key to the sub-optimality of debt restructuring within the program was the substantial expected cost of loss of market access and higher risk premia. Reprofiling could have reduced these costs (IMF 2015a), but not eliminated them, and would also have significantly reduced the gains. The EFSM loan reprofiling in 2013 crowded in the market rather than making market borrowing more expensive. A modality with these characteristics—i.e., some form of market-friendly official sector burden-sharing—would be consistent with the Fund’s recently redefined exceptional access policy for cases with significant risk of spillovers, and would be likely to lower total cost relative to unilateral default (especially taking spillovers into account). For more far-reaching proposals of burden-sharing, see Corsetti et al (2016), and—for a summary of recent proposals (red/blue bonds, a debt redemption fund, European safe bonds, etc.)—Brunnermeier et al (2016).

1 Edwards derives ‘appropriate’ haircuts for countries depending on their debt burden, severity of shock, etc., based on regression results from the universe of sovereign restructurings between 1978-2010.2 Debt subject to PSI—estimated using MLT debt excluding holdings of resident banks, social security, and the ECB’s SMP.3 Greece and Cyprus both saw milder consequences than the averages of these findings—returning to markets within about two years at rates of 5 percent. However, the consequences of these operations, which benefited from the low interest rate environment, were not known during the period relevant for PSI for Portugal.

51. Later, the EFSM maturities extension delivered important reprofiling benefits, though its reach was limited. It alleviated Portugal’s debt service burden without deterring other market participants—and indeed facilitated a EUR 6.6 billion bond exchange at end-2013 (with equal nominal value but refinancing deferred by 3 years). While it also illustrates the limits to rescheduling, since the face value of Portugal’s high debt stock remains intact, it demonstrated the way in which official reprofiling can crowd in new lending rather than igniting market fears of a further round. Particularly for countries in a currency union, where spillovers are a real concern, burden-sharing of this ring-fenced nature is likely to have significantly lower cost than unilateral debt restructuring and hence be a more tolerable approach to reducing debt stocks over time.

E. Was the Competitiveness Strategy Realistic?

52. Improvements in competitiveness were a main objective of the program, but the gains have been limited. In the absence of the exchange rate instrument, and with growth low for the decade before the crisis, it was appropriate for the program to make competitiveness a core objective, through internal devaluation and structural reforms. But despite a wide-ranging structural reform program, and compliance with practically all structural benchmarks, growth is still tepid and unemployment remains high. The rapid closure of the current account deficit should not be confused with the achievement of competitiveness, since Portugal is still far from internal balance. The ULC-REER is estimated to have declined by 7 percent, but the CPI-based REER to have declined only marginally. This supports the conclusion of the 2015 Crisis Program Review that internal devaluation is arduous and takes a long time, but also raises the question of whether the strategy could have been designed better.

53. The impact of an effective internal devaluation was difficult to see in the macro-framework. The strategy (for the short term) was to compress ULCs, but the program assumed only a marginal pass through to final prices. Growth was projected to resume in 2013, but the decline in the CPI-REER implicit in the program request was projected at only ½ percent by 2014, begging the question of how competitiveness gains would be compatible with growth. The cumulative price differential against the euro area was slightly over 1 percent and could not have been expected to go far in eliminating the price competitiveness gap. Also, while costing of reforms was attempted, it proved difficult to establish the link between reforms in competitiveness and their effects on prices. In the absence of a more developed toolkit, it has been prudent practice in Fund programs to assume that structural reforms will yield fruit in the medium term only. However, for programs where internal devaluation and competitiveness improvements are vital for sustainability, more specification of the mechanics of effective structural reform is called for, and the macro-framework should show how program objectives will get delivered.

uA01fig16

Cumulative CPI Inflation Differential from 2011: Portugal and Euro Area

(Percent)

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Sources: IMF, World Economic Outlook; and IMF staff calculations.

54. The structural reform program was broad, and benchmarks included measures only tenuously linked to relative price adjustment or growth. While the program contained 55 benchmarks (Table 3), the list of structural actions in the first MOU of European partners ran to about 27 pages. Gershenson et al (2016) estimate that a total of 494 different actions were taken (to initiate, legislate implement or evaluate structural reforms), about half in the public sector and a third in product markets. A more parsimonious approach, designed around well-targeted benchmarks, could have shed more light on the link between objectives and actions, and eased the implementation burden for the authorities.

55. Reforms targeting changes in income relativities or the loss of acquired rights had difficulty in getting ownership. The Constitutional Court reversed reforms to public sector compensation, on the grounds that the burden was not evenly shared across the population. Important reforms flexibilizing wage bargaining have also been reversed. Efforts to reduce rents in non-traded goods sectors met with little success in the face of vested interests. None of this is surprising, given the goal of the reforms, but it illustrates the difficulties in achieving effective internal devaluation.

56. Sequencing deficiencies may have reduced the effectiveness of reforms. In Portugal, labor market reforms might have been more effective if product markets had been liberalized. The Spring 2016 WEO (IMF 2016) finds product market reforms to generate short-term gains, while the effectiveness of labor market reforms depends on type of reform and varies with economic conditions. Blanchard and Giavazzi (2003) suggest that sequencing of competitiveness reforms matters: for savings in labor costs to pass through to prices, product market reforms dismantling excessive rents have to precede labor reforms.17 However, experience from transition economies shows that, in practice, the contentious political economy surrounding structural reforms makes them difficult to sequence: it is more pragmatic to undertake reforms when a political opportunity presents itself, with attention to avoiding the worst inconsistencies.

F. Should the Strategy for the Banking System Have Been More Proactive?

57. Financial sector problems remain. Since the program ended, asset quality has continued to deteriorate, the third largest bank (BES) failed soon after the program finished, and further recapitalization needs have been revealed in several banks. The program has been criticized by others for not having taken a more stringent approach to identifying banking system vulnerabilities, as prelude to a more aggressive strategy for strengthening bank capital and deleveraging corporations.

58. Portugal’s vulnerability diagnostics shared many characteristics with those in other euro area program countries… As in Ireland, Greece, and Cyprus, the assessment framework included independent loan reviews by qualified third parties based on a robust methodology, with comprehensive coverage, and under a well-designed governance framework. The loan reviews took an accounting (“book value”) perspective (i.e., “incurred loss approach”, consistent with IFRS); they were not backed by a real estate appraisal (as in other countries), but this was justified by lack of a preceding property bubble. Based on the information from these reviews, the ‘middle-ground’ strategy of keeping banks open but under review for adequate capital and a realistic degree of loss recognition was justified.

59. … but the forward-looking assessment was an area of difference. The stress test in Portugal was not based on an independent loan loss forecast as in other program countries.18 Instead, loan loss forecasts were modeled by banks consistently with the BdP’s credit risk model, based on common macro-scenarios. Different asset classification methods and risk parameter definitions across banks meant that consistency could not be guaranteed. The independent review of stress testing proposed improvements in banks’ stress tests, but these came into effect only after the initial SIP exercise. Also, in 2013, some inconsistencies were identified in the BdP’s stress testing framework that might have resulted in too optimistic projections. By end-program, the BdP had significantly revamped its stress-testing framework in collaboration with the Troika.

60. The somewhat different approach in Portugal was justified by conditions in the banking system at the time of program design. Portugal was not confronted with a banking crisis, and a program goal was to avoid opening another flank of instability, particularly at a time when fears of contagion to the rest of the euro area were high. The authorities argued that independent scrutiny comparable to that in crisis countries would weaken confidence in Portuguese bank soundness.

61. In hindsight, given the highly leveraged corporate sector, both the macro-framework and the bank vulnerability assessment should have been more conservative. In 2010, over 40 percent of corporate debt was at risk and about 40 percent of companies had an interest coverage ratio below 2, which is generally considered a worrisome threshold. This was a big latent problem, which may not have been sufficiently recognized in the design of the program. Deleveraging was necessary in all sectors (banks, private sector, government) and a firmer recognition of this would have inspired more conservative growth projections—and hence a more conservative diagnosis of asset quality and higher capital needs. Even in the absence of a full-fledged independent assessment, earlier identification of shortcomings in the risk model used by the BdP and more conservative calibration of the macroeconomic scenarios would have enabled a more realistic forward-looking assessment, with better identification of vulnerabilities and developing problems. There was significant room to increase the capital cushion against greater expected losses under the program, since less than half of the EUR 12 billion BSSF was drawn down.19

62. But even if banks had had larger up-front re-capitalization, this would have not been sufficient by itself to catalyze corporate debt restructuring. Larger capital buffers would have enabled banks to handle debt restructurings better. But to induce banks to take more BSSF capital, the terms (pricing) and conditions (mandatory downscaling) regulated by EU State Aid rules would have needed to be less of a deterrent. And even then, as demonstrated by Ireland’s experience, larger upfront capital with measures to facilitate corporate debt restructuring are still insufficient when growth is sluggish, since both banks and borrowers have weaker incentives to engage in restructurings. In such cases, a comprehensive approach is needed, to establish a legal and institutional foundation and confront incentive misalignments (see below).

63. It is unlikely that a more stringent asset quality assessment would have revealed problems in BES.20 BES was part of a large and complex mixed conglomerate (Espirito Santo Group) operating cross-border with an opaque structure and governance. The BdP was the supervisor of BES and its parent financial holding company, but in retrospect, lacked a full grasp of risks in the conglomerate and in bank operations of the Angolan subsidiary, one of the largest profit centers in the financial group. Such risks could have been addressed only through a more hands-on supervisory approach (including of global cross-border operations across multiple jurisdictions—which is particularly challenging), more recognition of risks from the group (stronger conglomerate supervision), and more preemptive actions, including to improve governance in the group. Finally, fraudulent behavior, as revealed by the ex-post forensic review commissioned by the BdP, is typically harder to uncover preemptively in such complicated legal and operational group structures.

64. A more thorough scrutiny of supervisory practices followed by additional measures to strengthen supervision would have benefited the program. The repeated fraud incidents in the Portuguese banking system (BPN, BES) raise the question of whether the BdP had a sufficiently robust approach to monitoring governance in banks. A deeper assessment of supervisory practices against the Basel Core Principles (BCP) for Effective Bank Supervision could have been useful to reveal weaknesses in supervision and propose remedial measures under the program.

65. Finally, it is important to recognize that the program ended before some current bank fragilities emerged. The weak economic, low interest-rate, environment has squeezed banks’ financial positions, so not all current bank problems are unfinished business from the program.21

G. How Could Corporate Restructuring Have Been More Effective?

66. Resolving systemic corporate over-indebtedness is generally too long-term and complex a task to be achieved within a program. A program could only initiate deleveraging. Market-based solutions are almost always preferable to government solutions, but work only when all stakeholders (e.g., banks, companies, government) have adequate incentives to engage in the process. Towards the end of the program, staff sought an overhaul of the debt restructuring strategy to better reflect the structure and needs of the corporate sector. But the impediments to effective debt restructuring in the absence of sufficient capital buffers were not adequately recognized.

67. With benefit of hindsight, a more comprehensive and proactive strategy to address the corporate debt overhang could have yielded better results (Box 7):

  • First, corporate vulnerabilities could have been more firmly reflected in banks’ asset quality assessments. More granular guidance on estimating incurred losses and a more conservative framework for estimating expected losses could have led to a more realistic estimate of the necessary capital buffers. Stronger capital buffers would have prepared banks better to respond to the deterioration in the corporate sector by more aggressive write-offs or restructurings. Also, banks should have been encouraged to repair their balance sheets through more aggressive write-offs or restructurings—for example, by strengthened supervisory overlays in loan provisioning and classification or by introducing quantitative restructuring targets by type of loan.

  • Second, a more targeted strategy for corporate debt restructuring could have been designed, tailored to company size. Past episodes of large-scale corporate debt restructuring point to the importance of proper segmentation of companies—to apply strategies appropriate to companies’ size and economic relevance. In Portugal, a strategy tailored to microenterprises and SMEs was a priority, given their large share of employment and gross value added.

H. Was Ownership Adequate?

68. At the inception of the program, there was strong ownership to deliver the necessary adjustment, but this declined over time. Broad political support was expressed in writing by the two main opposition parties. The transition to the new government—a time of extraordinary risk to an incipient program—was exceptionally well-handled. However, ownership dwindled in the face of recession and adjustment fatigue. Portugal’s exit from the program without completing the last review, and choice not to ask for a follow-up precautionary program, left an unfinished reform agenda.

Toward More Effective Corporate Debt Restructuring

Studies demonstrate that corporate debt becomes a drag on growth when it exceeds 90 percent of GDP (Cecchetti et al, 2010) and has substantial and persistent negative effects on investment (Jaeger, 2003 and Goretti and Souto, 2013). Risks from excessive debt get amplified in times of economic distress through a negative feedback loop: corporate defaults and NPLs erode bank capital and constrain credit.

Both Portugal and Ireland took important steps to promote deleveraging, but neither has succeeded in entrenching a transformative corporate debt restructuring process. An effective strategy probably needs to take a 360 degrees’ approach, to align incentives of all stakeholders (banks, borrowers, and government). Policies should respond to the specific conditions of banks and corporates,1 and should include:

  • An overarching strategy for corporate debt restructuring and a well-tailored restructuring framework, reflecting the structure and characteristics of the corporate sector (standardized restructuring agreements for SMEs, etc.); such a strategy could introduce sweeteners for companies to engage in debt restructurings (government guarantees) or measures to diversify firms’ funding strategies;

  • On the banks’ side: accurate recognition of risks (through more conservative asset classification and provisioning, improved disclosure of restructured loans, and strong supervision); capital buffers with forward-looking calibration to ensure that banks engage comfortably in restructuring; operational preparedness to deal with debt restructurings; and enhanced supervisory oversight;

  • Improvements in the legal infrastructure (bankruptcy laws) to facilitate smooth debt enforcement and rehabilitation of viable companies; and building capacity of the judiciary (judges, insolvency administrators) - to deal with corporate bankruptcies, but also with large scale debt restructurings;

  • Enhanced transparency and communication across all stakeholders.

  • Finally, better understanding of the link between corporate and bank balance sheets should be a central element of macro-financial analysis.

For SMEs, examples of restructuring programs are scarce,2 but suggest that standardized restructuring agreements, including specified grace periods for repayment and predetermined schedules for interest rate reductions, can yield good results. The standardized SME restructuring program would include the following features: a simplified screening of debtors for viability, a decentralized bank-led process based on harmonized restructuring terms (grace periods, capitalization of interest in arrears, extension of loan maturities), arbitration in case of disagreement between banks and debtors, the injection of new funds (working capital) in viable firms, 3 and regular reporting by the banks. To ensure progress in dealing with a large number of SMEs, banks may be given quarterly restructuring targets (monitored by the supervisor).

The largest firms typically require complex negotiations on an individual basis. Restructuring of large corporates requires private expertise—financial advisors, insolvency specialists, auditors and lawyers—to help preserve as much going-concern value as possible. Debt workouts combined with deep corporate restructuring are preferable and send the right signal to other distressed firms that may see debt restructuring as a bail-out. In such cases, the workouts should be facilitated via the introduction of a credible and efficient out-of- court restructuring regime. Such a regime would be a creditor-led workout process that facilitates debt restructuring via time-bound negotiations, provides for more effective inter-creditor coordination via creditor coordination committees, and include expedited arbitration procedures to settle any material differences of opinion, either among creditors or between debtors and creditors.

Finally, it is important to recognize that the economic environment also plays a significant role in boosting or discouraging debt restructuring. For example, Korea’s comprehensive debt restructuring after the Asian Crisis is considered a successful role model—but benefited from strong post crisis recovery, which improved incentives to engage in debt restructuring compared to those facing euro area crisis countries.

1 See, for instance, Aiyar et al (2015) and Bergthaler et al (2015).2 In Iceland, in response to sluggish restructurings (2010), the authorities opted for an across-the-board restructuring of SME debt, with calibrated debt relief from banks and strong incentives for SMEs to raise equity.3 The Portuguese Mutual Guarantee System has financed healthy SMEs, but has not been tied to SME workouts.

69. Several elements conspired to erode commitment to the program.

  • Slippages and the periodic announcement of new measures undermined faith in the program and contributed to austerity fatigue. This argues for avoiding over-ambition in a program: over-performance in delivering realistic targets is more likely to maintain loyalty.

  • The interpretation of the constitution as ruling out changes in relativities on grounds of fairness raised doubts about the legitimacy of the program; it also meant that the future became much less certain, since measures seemingly adopted could be reversed.

  • The frontloading of program financing meant declining incentives to see the program through to the end.

  • Portugal’s re-entry to markets allowed it to finance a looser fiscal policy.

  • Indirectly, the ECB ‘whatever it takes’ speech, by giving confidence of a better financing environment going forward, reduced banks’ incentives to restructure corporate debt.

70. The setbacks to the program from constitutional court rulings were an important problem. Not only did the reversals create uncertainty, but also the replacement measures were of lower quality (delivering fiscal adjustment, but less growth-friendly and inimical to gaining competitiveness). There are diverging views about whether they could have been addressed more effectively. Some analysts are of the view that better legal drafting (with more understanding of the legal constraints) would have made the measures acceptable. Others believe that, after two or three reversals, the Fund should have made completion of further reviews contingent on assurances of the legality of the measures. All acknowledge that, while any program is a compromise between the desirable and the feasible, the Fund may have to walk away if erosion in the quality of the package puts the goals of the program out of reach. In that case, analysts emphasized the importance of communications to explain to the public the impact of the legal framework.

Constitutional Court Reversals of Cost-Cutting and Flexibilization Measures

article image
Sources: The Portuguese Crisis and the IMF, Eichenbaum et al, IEO, 2016; The Economic Adjustment Program for Portugal, European Commission OP 202, 2014

71. The availability of ECB financing was a blessing and a bane. It is difficult to see how stabilization would have been achieved without the loosening of financing constraints. However, the perception of unconditional support (‘whatever it takes’) may have unavoidably created an element of moral hazard: markets were ready to finance looser fiscal policy and banks could delay pursuing loan restructuring. This underscores the importance of having a monetary policy strategy which is fully aligned with achieving the goals of the program. In cases where this is not possible (e.g., as in the case of the ECB, where monetary policy is designed to meet union-level goals rather than being tailored to the needs of a specific country), it raises the question of whether a Fund program can put adequate measures/incentives in place to ensure program goals can be delivered despite lack of control over the monetary environment.

72. When asked what the Fund could have done better to support ownership, analysts emphasized communications. The Troika was perceived as reticent, leaving the government to explain the rationale for the program and its amendments. A more effective defense of the program could have created a broader social understanding of the goals and costs of failure. However, Fund staff, in fact, did extensive outreach, and interlocutors had little appetite for more Troika presence in the media—meaning that probably what was needed was more emphasis in the program on maintaining a constructive government communications strategy.

Consistency With Fund Rules and practices

A. Were the Exceptional Access Criteria Observed?

73. The first criterion was fully met. Balance of payments pressures remained high, with large refinancing needs throughout the program.

74. Use of the systemic exemption in place of the debt sustainability (2nd) criterion was more problematic. The concern is not the exemption itself, since Greece and Ireland had previously applied it. However, remarkably little justification was provided (either in program papers or orally to the Board) about why Portugal risked triggering a systemic crisis. An estimation of possible contagion to other countries and pressure on exposed core banks was clearly warranted. Staff saw it as an appropriate judgment call under incomplete information, and this report does not disagree. But, looking forward, if risk of contagion were to be used again as a rationale for UFR, staff should base the case on more structured analytical arguments. Further, during the program the risk of contagion waned as foreign banks’ claims on Portugal shrank by about 30 percent (left panel in Figure 10). European firewalls were established, and the ECB clarified its support for euro area members). It is not obvious that spillover risks remained systemic during later reviews; staff analysis should have reconfirmed this to the Board. (That said, it is hard to imagine revoking eligibility for exceptional access late in a frontloaded program.)

Figure 10.
Figure 10.

Portugal’s Integration with Countries in the Euro Area

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

75. Portugal achieved a well-managed re-entry to markets in advance of the program target, thereby meeting the 3rd criterion. 22 However, spreads did not begin to decline significantly until the ECB announced euro-area wide actions (right panel in Figure 10),23 including its commitment to do “whatever it takes” to save the euro. Thereafter, Portugal’s access to markets was clearly facilitated by, if not completely dependent on, accommodative ECB financing.

76. Finally, the erosion of ownership over the program raises questions about whether the fourth criterion, reasonable prospects for program success, continued to be met. Given the substantial elements of success in the program, accepting the criterion was justified. But the justification became less solid as more measures were reversed. In the case of Portugal, and in general, the Fund would benefit from a more articulated approach to responding to lack of ownership—what assurances to ask for (including on the legality of the proposed measures?); how outspoken to be about the cost of compromise solutions; and when to walk away.

B. Was Prior Surveillance Adequate?

77. Pre-crisis staff reports persistently warned of the risks from fiscal and external imbalances and weak growth, but not of a sudden stop or adverse financial-real sector feedback. Risks from sizable external imbalances, weak productivity and declining external competitiveness were highlighted in the 2006-2009 Article IV reports, as was the need for expenditure-based fiscal consolidation to ensure medium-term sustainability.24 Areas that received less focus in pre-crisis surveillance were the country’s growing vulnerability to a sudden stop in capital flows, and emerging risks to the financial sector from sizable exposure to an unproductive corporate sector. As observed by other commentators, at the time the concept of sudden stop was not considered relevant to advanced countries. And Fund understanding of macro-financial linkages has been a work-in-progress over the past decade.

78. The 2006 FSAP and pre-crisis TA reports did warn of increasing risks for banks and corporates. The FSAP noted risks from high household and corporate indebtedness as well as substantial credit concentration and reliance on foreign funding, and recommended that these be monitored. In 2010, technical experts signaled that heightened tensions in the banking and corporate sectors, including considerable expected weakening of corporate debt servicing capacity, could lead to disorderly deleveraging. Staff recommended firmer policy measures, including encouraging banks to build capital and liquidity buffers aggressively, stepping-up supervision, and enhancing contingency planning and communication across agencies.

79. Surveillance could have focused more on the broader public sector and related fiscal risks. The 2003 Fiscal ROSC stressed that the confining the coverage of fiscal accounts to the general government might lead to oversights on public institutions. SOEs were used as off-budget vehicles for social policies, putting pressure on fiscal sustainability. In 2010, the deficit of non-financial SOEs amounted to ½ percent of GDP while their liabilities amounted to 36½ percent of GDP, and the treasury had to provide financing to avoid defaults, which generated a cash drain of about 3 percent of GDP during 2011. Most SOEs were not consolidated with the general government, and, while the program envisaged addressing their challenges, they were not covered by the floor on the consolidated general government. While the program strengthened the monitoring, management and disclosure of fiscal risks—including SOES, PPPs and other contingent liabilities—stronger pressure from Fund surveillance before the crisis might had helped with a better understanding of financing needs and limited damaging surprises from these sources.

C. Did the Troika Arrangement Inappropriately Constrain Fund Operations?

80. By the time of Portugal’s program, operating modalities with the Troika had been well-developed. There was praise from all stakeholders about the collaboration (rather than the frustration expressed in other cases), despite substantial burdens of coordination. As in other cases, the Fund was not perceived as a junior partner; in particular, Fund expertise in crisis management, reform program design, and technical issues was seen as valuable and difficult for other agencies to replicate.

81. There were, however, some objective effects on the program due to the jointness with EU partners. Namely:

  • Pace of fiscal adjustment. The fiscal path was determined by the EU’s requirement that the program show Portugal returning to compliance with SGP deficit targets. As discussed, the Fund would have preferred a more gradual path, and in the event, the target was postponed by a year.

  • Rigidities in financing. There was no effort to revisit the financing package when the fiscal targets were relaxed. The additional financing was manageable for the government, but the prospect of having to undergo a multicountry parliamentary approval process was clearly a deterrent to reconfiguring program financing—meaning that the Fund would have had to make a more awkward choice than usual if it determined that the program needed to be amended.25

  • Scope of structural program. The structural program was far broader than the Fund’s usual agenda. To some extent, this was justified by the centrality of the growth and competitiveness objective. But to some extent it reflected the broader EU vision of structural reform, including progress with meeting directives which only very indirectly contributed to macro-stability and sustainability.

  • Banking strategy. The ECB saw merit in a bigger bank solvency backstop and stricter deleveraging paths for banks, raising conceptual questions about whether it faced a conflict of interest in the program design given its exposure to the banks. As discussed above, a more proactive approach to bank recapitalization and to corporate debt restructuring could have served Portugal well.

  • A determinative role for ECB financing. Most importantly, the ECB’s accommodative monetary stance was determinative for program outcomes.26 Its financing was key to banks’ survival, although the low interest rate environment is now sapping profitability. On the other hand, Portugal’s debt sustainability will depend closely on the maintenance of low real rates for some time.

  • The ECB as monetary authority—not as Troika member—affects Fund operations. Portugal is a good example of the issue flagged in the 2015 Crisis Program Review, namely that the role of the regional monetary authority has important implications for program success. And hence that incorporation of ECB assurances into the program might have to be a prerequisite to being able to assure the Board that the program will be successful.27

Lessons

Overall assessment

82. This report concurs with other evaluations of Portugal’s program that it was a qualified success.28 Its main successes, which were of vital importance, come under the heading of successful crisis response. Macro-imbalances were closed—the current account deficit fully and the fiscal deficit significantly, despite reversals and a less-than-optimal quality of adjustment; international market access was regained; and a financial sector crisis was avoided (both in Portugal and elsewhere in the euro area). However, large stock disequilibria remain, with public debt high, and an unresolved private debt overhang leaving the banking system looking increasingly fragile and impeding corporate recovery. While some gains were made in structural reform and in competitiveness indicators, these have not been adequate to restore internal balance—meaning that growth remains subdued and unemployment high. The program delivered stabilization but not sustainability.

83. The report believes that the ‘big decisions’ taken in the program were justified. Namely: large upfront fiscal adjustment was the only practical strategy; debt restructuring was never a realistic option given the immediate costs and risks, and uncertain benefits; treating the banks as going concerns was justified by the information available; and labor market reforms were key to effective internal devaluation—even if the timing was less than ideal. In several instances discussed earlier, fine-tuning these decisions could have led to better outcomes, but it would be unrealistic to imagine holding up a crisis program to achieve an optimal composition of fiscal measures or correct sequencing of structural reforms. The accommodation of a looser fiscal path as the program progressed responded to concerns about growth.

84. In hindsight, some of the attention paid to protecting growth may have delayed inevitable adjustment. The strategy to go slow with deleveraging did not protect growth (although the counterfactual could have been worse). A more proactive strategy for bank reform and corporate restructuring could have put Portugal in a better position by now. But it would have required much bigger capital support (with implications for debt or fiscal adjustment), upfront acceptance of a deeper recession, and comprehensively resolving the impediments to corporate restructuring that have eluded other crisis countries.

Program duration

85. The implication of Portugal’s experience is that only so much can be done in a crisis program, even an EFF. When stock problems are large, or when internal balance must be restored by internal devaluation, a longer-term engagement may be needed to ensure a country’s sustainability (and safeguard Fund resources). Portugal’s experience supports the emerging consensus that internal devaluation can improve competitiveness, but is a long process, and recessionary. The report concurs with staff’s advice to government that a successor program for Portugal would have been helpful, to guide the unfinished adjustment process to solid completion.

Policy development needed

86. More policy development is needed to support countries undergoing internal devaluation. First, regarding the nature of Fund engagement: if longer program engagement would be helpful, would two successor EFFs be the best modality? And how to get country commitment to the longer process, when adjustment fatigue is hard to avoid even in current programs? Second, to tailor the financial program: not only to get growth and price projections right, but also to find the right balance between support (fairness) and adjustment for high-cost sectors being downsized. Third, with respect to the composition of the structural program: to identify and cost a more parsimonious set of reforms that efficiently release bottlenecks. Fourth, as regards sequencing: this report takes the view that, while sequencing matters and should be taken into account where possible, policymakers will rarely have the luxury of following an optimal sequence. Hence, policy development should focus on identifying the most problematic sequencing errors, so that they at least can be avoided.

87. Policy development—ongoing in the Fund—is also needed to help countries achieve more effective private sector deleveraging. Getting a not-too-hot/not-too-cold balance of bank and corporate deleveraging right is difficult: too fast a pace of bank deleveraging could have become disorderly for corporates, but the slow pace of corporate deleveraging in Portugal left much of the corporate sector frozen, slowing growth and damaging banks as impairments worsened. Experience in Portugal and Ireland suggests that facilitation measures are not enough to overcome banks’ and firms’ disincentives to take losses. A comprehensive plan of attack is needed.

88. Some rethinking of the treatment of the public sector in Fund programs and in statistics should also be considered. With general government as the statistical standard, public enterprises can move in and out of the perimeter of government depending on their profitability. While this may be appropriate from a financial control perspective, it may delay identification of worrisome trends in the public finances. Extending the perimeter from general government to the public sector would improve transparency and understanding of the role played by state enterprises in sovereign sustainability.

89. The report supports the decision to grant Portugal exceptional access, but sees shortcomings in applying the policy. While the systemic exemption was being appealed to, there was remarkably little analysis to justify it. The systemic exemption has been eliminated, but if in future the Fund is called on to commit billions of SDRs on an emergency basis to help prevent contagion, the analytical and evidential case for doing so should be more developed and more explicit.

Lessons relearned

90. Portugal’s experience supports lessons previously identified in evaluations in other crisis countries.

  • Over-ambitious targets are better avoided. Any gains from mobilizing exceptional effort are offset by social frustration and decline in program credibility when targets are missed and have to be recovered by additional measures. This means that estimated yields for measures should be pragmatic and underlying growth assumptions should be realistic.

  • Nonetheless, some frontloading of program measures is advisable, to minimize the impact of eroding ownership on outcomes. Getting the balance right between effective frontloading and self-defeating over-ambition is unavoidably a judgment call. Likewise, frontloading of financing can make a program credible at the most difficult moment, but the extent of frontloading is constrained by the need to preserve financial incentives to continue with the program through later reviews.

  • Debt restructuring gets harder, and is less effective, when not undertaken early. While there was never a good time for Portugal to undertake debt restructuring in the program, Box 6 shows that a haircut would have been largest if taken upfront, and the likelihood highest of restructuring being compatible with a return to markets “within the timeframe that Fund resources remain outstanding” (exceptional access criterion 3).

  • There is no substitute for strong bank supervision, which must be wary and keep pace with real sector developments. A full-fledged updated Basel Core Principles assessment would offer the best perspective on the quality of supervision, but would be resource-intensive at a time when country authorities are battling other problems. The Fund could develop simplified risk-based assessments to ascertain the quality of bank supervision—to be undertaken early enough in a program that any serious weaknesses could be addressed in good time.

When court rulings prevent adjustment

91. The Fund will not always be able to find a workable compromise if legal constraints are inimical to economic adjustment. While countries are entitled to whatever legal system they choose, delivery of a successful adjustment program requires the system to make the adjustment possible. Programs are always compromises but (a) the Fund should make clear the cost of adhering to laws which weaken the feasible quality of adjustment; (b) where necessary, should seek assurances on the legality of measures before going ahead with a program or review; 29 and (c) in the extreme case, may need to walk away if the compromises required by the legal system become inconsistent with success.

Programs with countries in currency unions

92. The program with Portugal shone light on the to-do list for designing effective programs with countries in currency unions. The authorities and European partners expressed clear approval of Fund involvement in Portugal’s program, but a sense that some wrinkles need to be worked out to make this type of program most effective.

  • The need for policy development on internal devaluation was discussed above.

  • Since adjustment takes longer, the duration of programs needs to be revisited, or at least clarified.

  • In turn, this will raise a separate question of how to get ownership for the longer period.

  • And since adjustment takes longer, financing needs are likely to be greater; this is a cost of membership. Identification of financing needs and their allocation between the Fund and Europe were handled smoothly in the case of Portugal, but clarification about cost-sharing expectations between the Fund and currency union could smooth the path to any future programs.

  • Modalities for burden-sharing of debt need to be clarified in advance, in cases where traditional debt restructuring is not seen by the currency union as in its interest. For euro area countries where systemic risk could still be serious, an abrupt restructuring could be disruptive, even with strengthened regional firewalls. The EFSM maturities extension was an extremely valuable alternative to debt restructuring in Portugal, both by creating near-term fiscal space and by improving market perceptions of Portugal’s capacity to manage its other debt. To avoid a future situation where currency-union members cannot seek Fund help because debt restructuring would be intolerable, it would seem important for the Fund to work with European partners preemptively (i.e., in a non-crisis environment) to develop an alternative orderly burden-sharing process with the market-friendly characteristics of the EFSM extension.

  • More clarity on union-level conditionality is needed. As noted in the Crisis Program Review (2015b), conditions on union-wide policies are consistent with the Fund’s Articles of Agreement. “Article V, Section 3(a) mandates the Fund to adopt policies for the use of its resources that will help members to resolve their balance of payments problems and ensure adequate safeguards for use of the Fund’s resources. This provision thus establishes the Fund’s inherent ability to call for the adoption of union-level measures where such measures are necessary for the success of a member’s Fund- supported program and/or to safeguard Fund resources.” The conditionality guidelines could usefully be amended to make this point and discuss practical implications.

  • Even if union-wide conditionality were to be mainstreamed, it would not address the dilemma of a monetary policy inappropriate for an asymmetrically-shocked member country. For the Fund to be able to give the normal assurances in a program, it needs to know either that monetary policy will be supportive or that adequate mechanisms exist for counterbalancing adverse monetary movements. An exploration of such support mechanisms with euro area authorities would seem called for, to clarify instruments for ensuring program goals can be met while monetary policy is dedicated to union-wide objectives.

Annex Figures and Tables

Figure 1.
Figure 1.

Portugal: Real Sector Developments

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Figure 2.
Figure 2.

Portugal: Labor Market Indicators and Price Indicators

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Figure 3.
Figure 3.

Portugal: External and Financial Sector Developments

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Note: Non consolidated data. Debt includes loans and debt securities.Source: Haver Analytics.
Figure 4.
Figure 4.

Portugal: Fiscal Sector Developments

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Figure 5.
Figure 5.

Portugal: Bank and Corporate Leveraging

Citation: IMF Staff Country Reports 2016, 302; 10.5089/9781475538403.002.A001

Table 1.

Portugal: Selected Economic Indicators, 2009-2014, Outturn Vs Program Request

article image
1 The EFF column reflects the projections used in the program request approved by the Board in May 2011; for 2009-10 it reflects preliminary actuals. There have been a number of revisions to macroeconomic data since the end of the program, including the re-basing of the national accounts and the shift from ESA-95 to ESA-2010 for fiscal reporting.Sources: Bank of Portugal; Ministry of Finance; National Statistics Office (INE); Eurostat; and IMF staff projections.
Table 2.

Portugal: Quantitative Performance Criteria

Billions of euros, unless otherwise specified

article image
Sources: IMS Staff Reports

Domestic arrears increased by €159 million in October and €74 million in November. They decreased by €301 million in December.

Table 3.

Portugal: Structural Conditionality

article image
article image
article image
article image
article image

PA is ‘Prefer Action’, M is ‘Met’, MOD is ‘Modified’