## _cr16301

## Source details

**Canonical URL:** [_cr16301](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16301.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16301.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16301.pdf.json)

---

### Projection Methodology
- Scope and data sources:
  - Projections cover 2015-2100 and draw on the 2015 United Nations (UN) projections for fertility, mortality and migration.
  - Pension and healthcare spending projections based on the October 2015 IMF Staff Discussion Note “The Fiscal Consequences of Shrinking Populations”.
  - Base-year data: 2012 OECD data for pension and healthcare spending; EC 2015 Ageing report spending profile used to incorporate reforms up to 2060 for pensions.
- Key assumptions and mechanics:
  - Baseline = UN main demographic scenario; alternative UN scenarios for lower fertility and higher longevity used for sensitivity.
  - Pension projections: 2012 OECD base; EC 2015 Ageing report profile applied up to 2060; beyond 2060 pensions driven purely by demographic factors.
  - Health projections: 2012 OECD base; incorporate “excess cost growth” (growth of health care spending per capita minus real GDP growth per capita, controlling for demographics). For Portugal, excess cost growth is assumed to go from the 1981-2007 mean in 2015 to the mean in the sample of advanced economies (0.7 percent) in 2050, and then converge to zero in 2100.
- Coverage limits and rationale:
  - Impact of ageing on long-term care, education, and unemployment benefits not estimated here.
  - Long horizon to 2100 used to present long-term fiscal implications and to permit sensitivity testing under alternative demographic scenarios.

### Demographic projections (UN baseline) — key figures and dynamics
- Population and dependency:
  - Portugal’s total population projected to shrink by about 29 percent over 2015-2100.
  - Euro area population projected to decline by about 9 percent over 2015-2100.
  - Portugal’s old-age dependency ratio: 32 in 2015; projected to reach 68 by 2100 (highest among euro area countries).
  - Euro area average old-age dependency ratio in 2015: 28.
- Fertility, mortality, migration:
  - UN forecasts: modest, gradual fertility rise in euro area but remaining below 2.1 children per female.
  - Portugal: fertility projected to decline over 2015-2025, then slowly rise to reach its 1995 level by 2045; by 2100 Portugal would have the second lowest fertility rate among euro area countries (after Germany).
  - Mortality improvements for elderly projected to slow; by 2100 Portugal’s mortality rates remain slightly above euro area average for adults (15-64) and elderly (65+).
  - Migration: UN assumption—net migration inflows to developed countries constant over 2015-2050, then decline to half the 2050 level by 2100. Under the baseline Portugal’s net migration rates remain negative short term, become positive from 2025, and settle above most comparators (except Spain) from 2040 onward.

### Fiscal impact under baseline — age-related spending and composition
- Age-related public spending (pensions + health):
  - Portugal: projected increase of 6.1 percentage points of GDP by 2050.
  - Euro area average: projected increase of 4.5 percentage points of GDP by 2050.
  - By 2100, age-related spending increases by 7.4 percentage points of GDP for both Portugal and the euro area as a whole.
  - Portugal’s age-related spending would exceed the euro average by about 6 percentage points of GDP in 2050, and 5 percentage points in 2100.
- Pensions:
  - Portugal: public pension spending projected to increase by 1 percentage point of GDP by 2035, then decline by about 1.5 percentage point in the longer run (recent reforms generate savings).
  - Euro area: pension spending projected to increase by +0.6 percentage point by 2035 and by +0.5 percentage point over the longer run.
  - Portugal’s public-pension-to-GDP ratio would exceed the euro area average by about 4 percentage points of GDP over most of the projection period.
- Health:
  - Portugal: health spending projected increase of 5.7 percentage points of GDP by 2050, and 8 percentage points by 2100.
  - By 2100, excess cost growth explains 3/4 of the total increase in health spending in Portugal.
  - Portugal would have the highest health spending-to-GDP ratio relative to comparators and well above the euro area average.

### Fiscal implications and present discounted value (PDV)
- PDV of projected increase in age-related spending:
  - Equivalent to "106 percent of today’s GDP between 2015 and 2050, and 186 percent of GDP between 2050 and 2100."
- PDV assumptions:
  - Interest rate-growth differential converges to "1 percent by 2025 from their current level of -0.8 for advanced countries," based on medians from the World Economic Outlook (IMF, 2015a).
- Risks:
  - Under the baseline, age-related spending pressures would make Portugal’s public debt path unsustainable without offsetting policies.
  - Main drivers: low fertility, higher longevity, and (to a lesser extent) migration outflows; health excess cost growth is the dominant contributor.

### Quantitative outcomes across scenarios (selected table-style outcomes preserved)
- Baseline (2015, 2050, 2100; change 2015-2050, 2050-2100, 2015-2050, 2050-2100):
  - Old-age dependency ratio1: 32, 66, 69
  - Pension spending: 15.6, 16.0, 15.0, 0.4, -1.0, 23-16
  - Health spending: 6.8, 12.5, 14.8, 5.7, 2.3, 83 203
  - Pension + health care spending: 22.4, 28.5, 29.8, 6.1, 1.3, 106186
- UN Low fertility scenario:
  - Old-age dependency ratio1: 32, 73, 97
  - Pension spending: 15.6, 17.7, 21.3, 2.1, 3.6, 34 121
  - Health spending: 6.8, 13.8, 19.6, 7.0, 5.9, 63 08
  - Pension + health care spending: 22.4, 31.5, 41.0, 9.1, 9.5, 130429
- Longevity risk scenario:
  - Old-age dependency ratio1: 32, 70, 77
  - Pension spending: 15.6, 17.0, 16.8, 1.4, -0.1, 35 18
  - Health spending: 6.8, 12.9, 15.9, 6.2, 3.0, 88 223
  - Pension + health care spending: 22.4, 29.9, 32.8, 7.5, 2.9, 123241
- Note: "1/ Population 65 and older as a share of the population 15 to 65."

### Differences versus EC 2015 Ageing Report (key quantifications)
- Base-year data difference: IMF uses 2012 OECD data; EC used 2013 country-provided data.
- Using UN demographic projections without controlling for pension reforms would bring "Portugal’s public pension spending about 10 percentage points of GDP above the baseline by 2100."
- IMF staff’s public pension spending projections would be "higher by 1.5 percent of GDP in 2060 relative to the 2015 Ageing report baseline scenario."
- Health spending: excluding excess cost growth makes IMF projections comparable with EC 2015 Ageing report ("-0.1 percentage point of GDP in 2060 in the IMF projections").
- Excess cost growth scenario: EC finds increase of "4.9 percentage points of GDP in 2060"; IMF staff projections foresee "5.1 percent of GDP in 2060 due to excess cost growth." Difference reflects Portugal-specific excess cost growth of 1.9 versus EC average sensitivity of 1.4.

### Policy options to mitigate fiscal impact — overview and priorities
- Policy options considered:
  - Boost fertility rates.
  - Reverse net migration outflows / attract working-age migrants.
  - Increase labor force participation of women and the elderly.
  - Revisit recent pension reforms to promote equity across generations and schemes.
  - Lower health spending excess cost growth.
- Recommended approach:
  - Multi-pronged strategy combining demographic, labor market, entitlement, and revenue/expenditure measures.
  - Generate higher potential growth and productivity to offset fiscal impact.
  - Define a policy mix that accounts for interactions between options to avoid suboptimal outcomes.
  - Short-term policy action urged to frontload adjustments and avoid reversals.

### Policies affecting demographics — boosting fertility and migration
- Boosting fertility:
  - Quantitative effect: limited medium-run impact; more sizable long-run reduction in age-related spending, especially for Portugal.
  - Portugal context and constraints:
    - Childcare participation: "45 percent in 2013"; pre-primary enrollment "89 percent in 2012."
    - Maternal employment: "73 percent of women with children aged 0-14."
    - Family size: "around half of all Portuguese families have only one child."
    - Portugal spends "1.4 percent of GDP in 2011" on family and children benefits (euro area average "2.6 percent of GDP").
    - Eligibility threshold for family cash benefits: "€8,803".
    - Enrollment gap in subsidized childcare between high- and low-income families: "around 40 percent".
  - Policy implications: cash benefits have limited fertility impact; expanding subsidized childcare access for low-income families and capacity are more promising.
  - Recent measures (2016): family allowances increased "by 2 to 3.5 percent based on income brackets"; bonuses for single-parent households increased "by 20 to 35 percent"; a personal income tax reform replaced the family quotient by a lump-sum deduction.
- Reversing net migration outflows:
  - Risk of continued outflows: large impact on age-related spending, potential to reduce growth potential and competitiveness.
  - Existing measures: Golden visa scheme (from 2012) — "about 2,788 residency permits granted by end-2015"; Non-habitual resident tax regime — "about 5,600 foreigners had benefited ... at end-2014."
  - Recommendations: structural reforms to boost competitiveness and employment prospects; engage diaspora and remove reintegration barriers; attract working-age migrants and ensure labor market integration (introductory programs, avoid further minimum wage increases that deter low-skilled hiring, active labor market policies).

### Labor force participation, pension reform and equity
- Participation effects and magnitude:
  - Female labor force participation (ages 25-64): 75.5 percent in 2015 (Portugal) vs. 71.1 percent (euro area).
  - Cutting the gender participation gap by half over 2015-2050 → saving of about 1 percentage point of GDP by 2050 and 2100.
  - Cutting elderly (55-64) participation gap by half → reduce age-related spending by about 1.5 percentage point of GDP.
- Pension reforms and incentives:
  - Legal retirement age increased to 66 and linked to life expectancy at 65; sustainability factor will increase legal retirement age by more than two years by 2050 (authorities’ projections).
  - Pension coverage ratio expected to decline by more than 20 percentage points by 2035.
  - Authorities should avoid reversing adjustment mechanism; additional reforms needed to restore fiscal sustainability, increase equity, minimize informality incentives, and protect vulnerable groups.
  - Current reforms have protected current retirees: only about 15 percent of adjustment borne by those receiving pensions over next 20 years (median EU ≈ 46 percent); legal retirement age increase leads to about 80 percent adjustment in benefit eligibility for new pensioners within next two decades.

### Health spending — drivers, limits to private financing, and efficiency gains
- Composition and private financing:
  - Public spending on health below euro area average; total health spending higher (+0.3 percent of GDP in 2013) because private share higher.
  - Private health insurance coverage: 21.1 percent in Portugal vs. 46.3 percent euro area average.
  - Out-of-pocket spending increased; copayment exemptions progressive and not seen to impair access.
  - Public pharmaceuticals spending reduced to 0.4 percent of GDP in 2014 (euro area 1.2 percent).
  - Public wage bill in health sector still largely above EU average.
- Suggested efficiency gains:
  - Increase performance financial incentives in hospitals and primary care.
  - Increase share of generics.
  - Reduce length of stay in hospitals.
  - Expand nursing roles to reduce number of physicians without quality loss.

### Banking sector: drivers of profitability and vulnerabilities
- Pre-crisis shifts:
  - Net interest margins compressed; banks offset via alternative income and cost cuts (staff costs relative to assets fell from 1.3 percent in 1999 to 0.9 percent in 2007).
  - Regulatory capital to RWA declined to 7 percent at end-2007.
  - Loan-to-deposit ratio rose from 100 percent in 1999 to 150 percent in 2007.
  - Banks’ total assets rose from 184 percent of GDP in 1997 to 250 percent in 2007.
  - Loans to real estate and construction rose to 38 percent of NFC loans in 2007 (from <13 percent in 1997); construction loans rose from €8.4 billion in 1999 to €26.1 billion in 2008.
- Crisis episode (2008–2013):
  - Sudden stop in capital flows, liquidity stress, higher funding costs, reliance on ECB funding.
  - NPLs nearly tripled from <4 percent in late-2008 to 11 percent by 2013.
  - Net interest margins fell from 1.9 percent in 2008 to 1.1 percent in 2013.
  - System losses each year 2011–2014.
- Empirical findings (2010–15):
  - Portuguese banks: lower capital ratios than most euro area banks (except Greece); highest operating costs to assets ratio (with Italy and Greece); lower net income than most peers (except Greece, Spain, Ireland).
  - Econometric results: bank size and current capital positively impact profitability; real GDP growth significant macro determinant; public debt negative effect. In stressed economies capital impact larger, trading reduces profitability, domestic public debt has larger negative effect, EONIA positive.
- Public support and ownership changes:
  - Public injections: €7.2 billion (mainly €4.3 billion contingent convertible debt via BSSF).
  - Ownership shifts: BPI effectively controlled by CaixaBank and Angolan investors; BCP controlled by Sonangol (19.4 percent).
- Post-crisis (2014–16) constraints:
  - Legacy issues persist: NFC debt rose to over 120 percent of GDP during the crisis; remained 109 percent at end-2015.
  - Bank capital among lowest in EU; NPLs continued to rise.
  - Bank resolutions: BES (August 2014) and Banif (December 2015); fiscal costs: €4.9 billion for BES and €2.2 billion for Banif.

### Banks’ capital needs, regulatory environment, and APP/QE context
- Capital and business-model adjustment:
  - Significant loss-absorption capacity required; high cost of capital limits new equity injections.
  - Banks need to restore profitability via alternative stable income streams and cost cutting; capital likely needs to surpass regulatory minima.
- Regulatory short- vs medium-term effects:
  - BRRD, SRM, and future MREL impose short-term costs but medium-term benefits (improved asset quality and reallocation of credit).
- ECB Asset Purchase Program (APP) / QE:
  - APP initially €60 billion monthly to end-September 2016, extended to March 2017; monthly purchases rose to €80 billion from April 2016.
  - PSPP announced January 2015; CSPP launched June 2016.
  - Average monthly purchases of Portuguese government bonds: €1.2 billion (March 2015–June 2016).
  - ECB holdings of Portuguese government bonds: €19.1 billion at end-June 2016 (17 percent of outstanding, weighted average maturity 10 years).
  - Comparison: €19.1 billion ECB holdings vs. €26.4 billion IMF funding under the 3-year adjustment program.
  - Banks issued €1.8 billion covered bonds and €0.7 billion asset-backed securities since late-2014 programs; long-term LTRO/TLTRO loans to banks: €22 billion as of June 30.
- Transmission channels and observed effects:
  - Channels: portfolio balance, credit, wealth, confidence, exchange rate.
  - Fiscal/funding impacts: net bond issuance jump to €12.9 billion prior to June 2016, almost fully covered by ECB purchases; enabled €10.3 billion advanced repurchases to the Fund in 2015-16.
  - Ten-year yields fell by 300 basis points during 2014; spreads over German bunds decreased by 200 basis points; yields stable close to 3 percent in 2015-16.
  - Bank lending impacts: interest rates on new loans to NFCs < €1 million and new mortgages declined by 116 and 101 basis points (Dec 2014–May 2016). New loans to households rose 30 percent (y/y) in first half 2016; new lending to NFCs decreased by 5 percent in first half 2015 and by 18 percent in 2015.
  - NPLs: NFC sector 19.8 percent at end-2015; household sector relatively stable at 7-8 percent; mortgage performance relatively better.
  - Risks: limited space to reduce deposit rates further; a 100 basis point fall in Euribor estimated to reduce net interest income by ≈ €700 million (Bank of Portugal estimate).

### QE impact on consumption, investment, and macro composition
- Consumption and GDP composition:
  - Headline growth in line with pre-QE projections, but composition skewed toward consumption.
  - Consumption pickup largely absorbed in higher imports (openness and high import content of consumption).
  - Investment pickup modest; corporate debt overhang and contraction in bank lending to NFCs constrain investment.
- Portugal: contributions to year-on-year growth, 2014-15 (percentage points, unless indicated otherwise)
  - Total domestic demand: 0.5 4.6 4.1
  - Final consumption expenditure: -0.3 3.2 3.5
  - Gross capital formation: 0.8 1.4 0.6
  - Foreign balance: 1.7 -2.3 -4.0
  - Real GDP growth, percent: 2.3 2.4 0.1
- Confidence, wealth and real estate effects:
  - QE and preceding ECB signals contributed to recovery in consumer confidence.
  - Household consumption rose 6 percent while disposable income rose 1 percent (2013Q1–2015Q3); wealth effect from house prices important (1-percent house price rise ≈ ½-percentage-point decline in saving rate).
  - Portugal: 5.9 million housing units valued at €300 billion; house prices rose 9.9 percent between 2013Q1 and 2015Q4; estimated real estate asset value increase €25 billion; household savings decline €5 billion.
  - Financial asset side: estimated loss of €0.4 billion for a 50-bps decline in interest rates and €0.6 billion for a 100-bps decline.
- Macro impact and limits:
  - APP boosted consumption via confidence and wealth channels but impact on headline growth modest and expected to fade as savings normalize.
  - Limited transmission to corporate lending and investment due to corporate debt overhang and structural obstacles.
- Fiscal side-effects:
  - PSPP coverage of gross bond issuance reduced urgency of fiscal adjustment; structural primary adjustment in 2015-16 projected nearly 2½ percent of GDP less than pre-QE envisaged path.

### Policy implications and recommended actions (integrated)
- Fiscal and entitlement policy:
  - Refrain from pension policy reversals; revisit recent pension reforms to address equity concerns and to frontload adjustments to contain near-term spending pressure.
  - Combine pension and health spending restraint, rationalization of other expenditure, and/or increases in tax revenue as needed.
- Health sector:
  - Pursue efficiency gains forcefully (performance incentives, generics, shorter hospital stays, expanded nursing roles) to rein in excess cost growth.
- Labor market and demographics:
  - Increase labor force participation (women and elderly) and pursue migration policies to attract working-age migrants; reduce barriers to return migration and improve integration programs.
  - Strengthen growth potential via structural reforms to improve employment prospects and competitiveness.
- Banking sector:
  - Banks must bolster loss-absorption capacity and strengthen capital (likely beyond regulatory minima), reduce operating costs, and develop alternative stable income streams.
  - Address high NPLs, especially corporate sector, to unlock monetary easing transmission to corporate lending and investment.
- Timing and social consensus:
  - Policy action is urgent given recent population decline and spending pressures.
  - Pension structural reform requires clearly articulated social consensus involving all stakeholders to balance sustainability and equity.

*Source: _cr16301 — IMF staff report excerpts and calculations using UN 2015, 2012 OECD data, EC 2015 Ageing Report, and IMF analysis.*

### 1. Projection Methodology __________________________________________________________ 8

### 1. Projection Methodology

### Scope and data sources
- Projections cover the period 2015-2100 and draw on the 2015 United Nations (UN) projections for fertility, mortality and migration.
- Pension and healthcare spending projections are based on the October 2015 IMF Staff Discussion Note (SDN) “The Fiscal Consequences of Shrinking Populations”.
- Base-year data: 2012 OECD data for pension and healthcare spending; EC 2015 Ageing report spending profile used to incorporate reforms up to 2060 for pensions.

### Key projection assumptions and mechanics
- Baseline corresponds to the main UN demographic scenario; alternative UN scenarios for lower fertility and higher longevity are used for sensitivity analysis.
- Pension projections:
  - Use 2012 OECD as base year.
  - EC 2015 Ageing report spending profile is applied to capture the impact of pension reforms up to 2060.
  - Beyond 2060, pension spending is driven purely by demographic factors.
- Health projections:
  - Based on 2012 OECD data.
  - Incorporate “excess cost growth”, defined as the difference between growth of health care spending per capita and real GDP growth per capita, after controlling for demographic effects.
  - For Portugal, excess cost growth is assumed to go from the 1981-2007 mean in 2015 to the mean in the sample of advanced economies (0.7 percent) in 2050, and then converge to zero in 2100.
  - Excess cost growth captures rising health costs related to technology, the Baumol effect, and health policies and institutional settings.

### Coverage limits
- The impact of ageing on long-term care spending, education spending and unemployment benefits is not estimated in this paper.
- The European Commission 2015 Ageing report projects an increase of 0.5 percentage point of GDP in long-term care spending from 2015-2060 in its baseline, with alternative scenarios up to 2.5 percentage points of GDP.
- The OECD (2013) estimates long-term care spending increases between 0.8 and 1.3 percentage points of GDP from 2010-2060 under various scenarios.

### Rationale for long horizon
- Projections to 2100 are used to present the fiscal implications of demographic trends that will be fully realized only over the longer term.
- The methodology recognizes large uncertainties in long-term demographic projections and therefore uses alternative demographic scenarios for sensitivity testing.

---

### Findings: Demographic projections (UN baseline)
- Portugal’s total population is projected to shrink by about 29 percent over 2015-2100.
- Euro area population is projected to decline by about 9 percent over 2015-2100.
- Portugal would be among the euro area countries most adversely affected; some others (Croatia, Estonia, Greece, Latvia, Lithuania, and Slovakia) face even larger population declines.
- Old-age dependency ratio specifics:
  - Portugal’s old-age dependency ratio in 2015: 32.
  - Euro area average old-age dependency ratio in 2015: 28.
  - By 2100, Portugal’s old-age dependency ratio would reach 68, the highest among euro area countries.
- Fertility and mortality:
  - UN forecasts assume modest and gradual increase in fertility rates on average in the euro area but remaining below 2.1 children per female.
  - Portugal’s fertility rates are projected to decline over 2015-2025, then slowly rise to reach its 1995 level by 2045; by 2100 Portugal would have the second lowest fertility rate among euro area countries (after Germany).
  - Mortality improvements for the elderly are projected to slow over the projection period; by 2100 Portugal’s mortality rates would remain slightly above the euro area average for adults (15-64 years old) and the elderly (65 and over).
- Migration:
  - Portugal experienced three episodes of negative net migration historically; the baseline assumes a gradual normalization over the projection period.
  - UN assumption: net migration inflows to developed countries remain constant over 2015-2050, then decline to half the 2050 level by 2100.
  - Under the baseline, Portugal’s net migration rates remain negative in the short term, become positive from 2025, and settle above most comparators (except Spain) from 2040 onward.

---

### Findings: Fiscal impact under baseline
- Age-related public spending (pensions + health) projections:
  - Portugal: projected increase of 6.1 percentage points of GDP by 2050.
  - Euro area average: projected increase of 4.5 percentage points of GDP by 2050.
  - By 2100, age-related spending increases by 7.4 percentage points of GDP for both Portugal and the euro area as a whole.
  - Portugal starts from a higher level of age-related spending; as a result its age-related spending would exceed the euro average by about 6 percentage points of GDP in 2050, and 5 percentage points in 2100.
- Pensions:
  - Portugal: public pension spending projected to increase by 1 percentage point of GDP by 2035, then decline by about 1.5 percentage point in the longer run as recent reforms generate savings.
  - Euro area: pension spending projected to increase by +0.6 percentage point by 2035 and by +0.5 percentage point over the longer run.
  - Portugal’s public-pension-to-GDP ratio would exceed the euro area average by about 4 percentage points of GDP over most of the projection period.
- Health:
  - Portugal: health spending projected increase of 5.7 percentage points of GDP by 2050, and 8 percentage points by 2100.
  - By 2100, excess cost growth explains 3/4 (three quarters) of the total increase in health spending in Portugal.
  - Portugal would have the highest health spending-to-GDP ratio relative to comparators and well above the euro area average.
  - The increase in health spending is attributed mainly to cost-driven factors (excess cost growth, technology, Baumol effect, policies) rather than expanded coverage (Portugal achieved public universal coverage by the end of the 1970s).

---

### Fiscal implications and risks
- Under the baseline, age-related spending pressures would make Portugal’s public debt path unsustainable in the absence of offsetting policies.
- Main drivers of fiscal pressure: low fertility rates, higher longevity, and to a much lesser extent migration outflows; health excess cost growth is the dominant contributor to projected age-related spending increases.
- Long-term demographic projections are subject to large uncertainties; past UN projections have underestimated declines in fertility and mortality due to faster-than-anticipated demographic transitions.

---

### Policy considerations and mitigation options
- A multi-pronged approach is recommended to mitigate age-related spending pressure and preserve fiscal sustainability:
  - Policies affecting demographics and labor markets (e.g., measures to reverse migration outflows, increase labor force participation).
  - Entitlement reforms (avoid pension policy reversals; revisit recent pension reforms to address inequities among generations and across schemes).
  - Measures to rein in pension and health spending, rationalize other expenditure, increase tax revenue, or a combination of these.
  - Reduce healthcare cost growth by addressing excess cost growth drivers (technology, inefficiencies, institutional settings).
- Generating higher potential growth and productivity is key to offset the fiscal impact of a shrinking population.
- The respective mix of policies will depend on social preferences; a comprehensive mitigation strategy is required to safeguard social cohesion and fiscal sustainability.

---

*Source: _cr16301 - 1. Projection Methodology (August 25, 2016), IMF staff calculations using UN 2015 and 2012 OECD data.*

### 9.      Population aging in Portugal, if not offset, could put public debt on an unsustainable

### _cr16301 - 9.      Population aging in Portugal, if not offset, could put public debt on an unsustainable path

### Fiscal impact and present discounted value (PDV)
- The increase in age-related spending in Portugal is equivalent in present value terms to "106 percent of today’s GDP between 2015 and 2050, and 186 percent of GDP between 2050 and 2100."
- The PDV is estimated assuming an interest rate-growth differential that converges to "1 percent by 2025 from their current level of -0.8 for advanced countries," based on medians from the World Economic Outlook (IMF, 2015a).
- The paper does not estimate the impact of demographic changes on economic growth, nor the impact of lower growth on fiscal sustainability.

### Baseline and alternative demographic scenarios (selected quantitative outcomes)
- Baseline (2015, 2050, 2100; change 2015-2050, 2050-2100, 2015-2050, 2050-2100):
  - Old-age dependency ratio1: 32, 66, 69
  - Pension spending: 15.6, 16.0, 15.0, 0.4, -1.0, 23-16
  - Health spending: 6.8, 12.5, 14.8, 5.7, 2.3, 83 203
  - Pension + health care spending: 22.4, 28.5, 29.8, 6.1, 1.3, 106186
- UN Low fertility scenario:
  - Old-age dependency ratio1: 32, 73, 97
  - Pension spending: 15.6, 17.7, 21.3, 2.1, 3.6, 34 121
  - Health spending: 6.8, 13.8, 19.6, 7.0, 5.9, 63 08
  - Pension + health care spending: 22.4, 31.5, 41.0, 9.1, 9.5, 130429
- Longevity risk scenario:
  - Old-age dependency ratio1: 32, 70, 77
  - Pension spending: 15.6, 17.0, 16.8, 1.4, -0.1, 35 18
  - Health spending: 6.8, 12.9, 15.9, 6.2, 3.0, 88 223
  - Pension + health care spending: 22.4, 29.9, 32.8, 7.5, 2.9, 123241
- Notes:
  - "1/ Population 65 and older as a share of the population 15 to 65."
  - Alternative UN demographic scenarios estimate an even larger impact of demographics on spending pressure in the medium and long run.
  - A lower fertility scenario would "increase dramatically age-related spending in the long run and would catapult public debt ratios over the next 85 years, relative to the baseline."
  - The increase in both the old-age dependency ratio and in age-related spending would be larger under both alternative scenarios than for the euro area average and comparators.

### Differences versus the European Commission 2015 Ageing Report (key quantifications)
- Base data difference: this paper uses 2012 OECD data for public health and pension spending; the EC 2015 Ageing report used 2013 country-provided data.
- Using the UN demographic projections without controlling for pension reforms would bring "Portugal’s public pension spending about 10 percentage points of GDP above the baseline by 2100."
- IMF staff’s public pension spending projections would be "higher by 1.5 percent of GDP in 2060 relative to the 2015 Ageing report baseline scenario."
- Health spending: excluding excess cost growth from the baseline makes IMF projections comparable with the 2015 Ageing report ("-0.1 percentage point of GDP in 2060 in the IMF projections").
- Excess cost growth scenario: EC 2015 Ageing report finds Portugal would experience an increase of "4.9 percentage points of GDP in 2060"; IMF staff projections foresee a comparable increase of "5.1 percent of GDP in 2060 due to excess cost growth." The slight difference reflects use of "Portugal-specific excess cost growth of 1.9" versus the EC average cost sensitivity of 1.4.

### Policy options to mitigate fiscal impact (overview)
- Policy options considered:
  - Boosting fertility rates.
  - Reversing net migration flows.
  - Increasing labor force participation of women and of the elderly.
  - Revisiting recent pension reforms to promote equity among current pensioners and generations.
  - Lowering health spending excess cost growth.
- The paper emphasizes a "multi-pronged approach" that:
  - Partially addresses demographic challenges.
  - Increases labor force participation.
  - Reforms age-related spending programs.
  - Requires defining a policy mix that accounts for interactions between policy options to avoid suboptimal outcomes.

### Decomposition methodology (Box 3 — key assumptions)
- Pension expenditure to GDP (PE/GDP) is modeled as the product of:
  - Benefit generosity (average pension to GDP per worker).
  - Pension coverage (number of pensioners to population 65 and older).
  - The inverse of the labor force participation rate (population ages 15–64 to workers).
  - Aging represented by the old-age dependency ratio.
- Sensitivity tests assume constant benefit and coverage ratios, constant labor force participation and GDP per worker, so "pension spending as a share of GDP only varies with the old-age dependency ratio."
- A "no tax policy change scenario" is assumed in which tax revenues remain stable as a share of GDP.
- Health care expenditure to GDP (HE/GDP) is modeled as the product of:
  - Generosity of the health package for the young.
  - The inverse of the labor force participation rate of population 0–64.
  - A function depending on the ratio α (per capita health spending for older population to young; "3.2 on average for OECD economies") and the old-age dependency ratio.
- Under these health assumptions, changes in health care spending are a function of the change in the old-age dependency ratio, assuming the first two factors remain constant.

### Policies that affect demographics — Boosting fertility
- Quantitative effect:
  - A higher fertility rate would reduce the old-age dependency ratio more in Portugal than in comparator countries and the euro area.
  - In the medium run, the impact on age-related spending would be limited because fertility increases are gradual and the impact on the working-age population takes time.
  - In the long run, the reduction in age-related spending would be more sizable, particularly in Portugal.
- Policy context and constraints:
  - Portugal: participation rates of young children in childcare "45 percent in 2013" and pre-primary education "89 percent in 2012."
  - Maternal employment: "73 percent of women with children aged 0-14."
  - Family size: "around half of all Portuguese families have only one child."
  - Portugal spends "1.4 percent of GDP in 2011" on family and children benefits, services and tax measures (euro area average "2.6 percent of GDP").
  - Eligibility threshold for family cash benefits: "€8,803" (above the minimum wage).
  - Enrollment gap in subsidized childcare between high- and low-income families: "around 40 percent" (highest among OECD countries).
- Policy implications:
  - Cash benefits have limited impact on fertility; empirical studies show limited effectiveness.
  - Increasing childcare services, expanding capacity in subsidized childcare, and improving access for low-income families are highlighted as more promising measures.
  - Portugal increased progressivity of family support in 2016: family allowances increased "by 2 to 3.5 percent based on income brackets," bonuses for single-parent households increased "by 20 to 35 percent," and a personal income tax reform replaced the family quotient by a lump-sum deduction (measures aimed at reducing inequality and poverty but with limited expected impact on fertility).

### Policies that affect demographics — Reversing net migration outflows
- Risk scenario:
  - Continuation of recent migration outflows would have a large impact on age-related spending, translating into a sharp increase in the old-age dependency ratio and spikes in public pension and health spending.
  - A sustained period of large migration outflows could also lower growth potential by reducing labor force size and productivity and worsen competitiveness.
- Existing measures and outcomes:
  - Golden visa scheme (from 2012) for non-EU private investors: "about 2,788 residency permits granted by end-2015."
  - Non-habitual resident tax regime for highly-skilled foreigners and foreign pensioners: "about 5,600 foreigners had benefited from this tax regime at end-2014."
  - Attracting EU retirees helps indirectly but does not improve demographic structure.
- Policy recommendations to reverse outflows and increase inflows:
  - Increase potential growth and employment prospects to reduce migration outflows (structural reforms to boost competitiveness, prudent minimum wage policy, more inclusive and transparent social dialogue).
  - Engage with the Portuguese diaspora and remove barriers to reintegration to facilitate return migration.
  - Attract working-age migrants while ensuring successful labor market integration:
    - Introductory programs to support migrants in getting jobs.
    - Preventing further increases in the minimum wage to reduce barriers for low-skilled workers.
    - Active labor market policies to support private sector employers in hiring migrants.

*Source: IMF staff calculations.*

### 20.      Improving labor force participation of women and the elderly would contribute more

### 20.      Improving labor force participation of women and the elderly would contribute more

### Labor force participation: effects and magnitude
- Increasing labor force participation raises GDP by increasing the number of workers, assuming productivity remains constant.
- Higher labor force participation of women and the elderly generates limited long-run savings relative to an increase in fertility or an improvement in net migration.
- Female labor force participation for ages 25-64 was 75.5 percent in 2015 in Portugal, compared with the euro area average of 71.1 percent.
- Cutting the gap between men and women in labor force participation by half over 2015-2050 would generate a saving of about 1 percentage point of GDP by 2050 and 2100.
- For the elderly (aged 55-64), the labor force participation rate is comparable to the euro area average, but the gap with individuals 25-64 is larger.
- Cutting the elderly participation gap by half would reduce age-related spending by about 1.5 percentage point of GDP.

### Recent tax and pension reforms to raise participation
- Recent reforms include allowing individual income taxation to reduce the secondary earner tax wedge, to which labor supply is more responsive (IMF, 2013b).
- The legal retirement age has been increased to 66 years old, and is adjusted in line with the evolution of life expectancy at age 65.

### Incentives for older workers
- Older workers (aged 40 and above) have longer unemployment benefits than young workers, as they are eligible to additional social unemployment after the end of their unemployment insurance benefits (IMF, 2013a; OECD, 2015).
- Gradually aligning the benefit duration of older workers on young workers as employment conditions improve could increase work incentives for the elderly.

### Pension reforms: sustainability and equity
- The sustainability factor links the legal retirement age to life expectancy at the time of retirement, and reduces pension benefits for early retirement in line with improvements in life expectancy.
- The sustainability factor will increase the legal retirement age by more than two years by 2050, according to the authorities’ projections.
- As a consequence, the pension coverage ratio is expected to decline by more than 20 percentage points by 2035.
- The authorities should avoid reversing this adjustment mechanism, which is expected to contribute to reining in pension spending.
- Additional pension reform measures are needed to bring down spending in the more medium term to:
  - restore fiscal sustainability;
  - increase equity (between the workforce, pensioners and across generations);
  - minimize incentives for informality and inactivity;
  - retain protection for the most vulnerable.
- Pursuant to the Constitutional Court ruling on pension reforms, a comprehensive structural reform of the pension system is required to balance sustainability and equity among generations and across pension schemes (EC, 2014). This reform would require a clearly articulated social consensus involving all stakeholders.
- Recent reforms have protected current retirees while placing a substantial adjustment burden on the future generation of retirees (IMF, 2012):
  - Only about 15 percent of the adjustment will be borne by those who receive pensions over the next 20 years, compared to a median of about 46 percent of the adjustment burden in the EU.
  - The relatively rapid increase in the legal retirement age will lead to an adjustment of about 80 percent in the benefit eligibility within the next two decades for new pensioners, somewhat closer to the European median.
- A comprehensive reform should address inequities among generations (including current pensioners largely preserved from recent reforms) and across pension schemes, as benefit generosity for current pensioners varies significantly.

### Health spending: composition, limits to private financing, and efficiency gains
- Public spending on health in Portugal is below the euro area average, but total health spending is higher (+0.3 percent of GDP in 2013) because the share of private spending is higher than in most countries (except Greece).
- Coverage by private health insurance is low in Portugal: 21.1 percent, compared to 46.3 percent on average in the euro area, leaving a large amount of health spending outside public and private insurance coverage.
- Out-of-pocket spending for households has increased more in Portugal over the recent period, but large social exemptions have made copayments progressive and have not adversely impacted access (OECD, 2015).
- Recent efforts have focused on reducing public spending on pharmaceuticals (0.4 percent of GDP in 2014), which are now significantly lower than in the euro area (1.2 percent of GDP), and as a share of total public health spending.
- Portugal’s public wage bill in the health sector is still largely above EU average.
- Suggested efficiency gains:
  - increasing performance financial incentives in both hospitals and primary care;
  - further increasing the share of generics;
  - making the production mix less cost intensive by reducing the length of stay in hospitals;
  - expanding nursing roles to reduce the number of physicians without diminishing quality of service (OECD, 2015).

### Projected impacts and policy priorities (conclusions)
- Under the baseline, age-related public spending would increase by more than 7 percentage points of GDP, and public debt would become unsustainable. This would be driven by health spending, as population ageing and technological improvements are projected to increase faster than economic growth.
- Pension spending will further rise in the short term, as recent reforms have protected current retirees while postponing sharp adjustment on future generations of pensioners.
- Addressing the fiscal challenge requires policies to:
  - raise growth potential,
  - improve demographics and labor markets,
  - reform entitlements.
- Offsetting rising fiscal costs will require either:
  - reining in pension and health spending,
  - rationalizing other expenditure,
  - increasing tax revenue,
  - or a combination of these options.
- Policy action is needed in the short term because Portugal has already experienced a recent population decline; the pace of entitlement reform should be faster and policy reversals should be avoided.
- There is urgency to revisit recent pension reforms to frontload the adjustment effort to:
  - contain spending pressure in the short run;
  - ensure more equity between current pensioners and future generations of pensioners.
- Reforms to reduce health care cost should be pursued forcefully to ensure fiscal sustainability of health costs.

*Source: IMF staff calculations using UN 2015, OECD, EC 2015 Ageing Report, and IMF country analysis contained in the chapter.*

### 2.       The drivers of bank profitability

### 2.       The drivers of bank profitability

### Change in profit drivers (pre-crisis)
- Net interest margins (NIM) were increasingly compressed by competition following euro area accession.
- Banks preserved profitability by strengthening alternative income sources and decreasing costs:
  - Staff costs relative to assets fell from 1.3 percent of assets in 1999 to 0.9 percent of assets in 2007.
  - Cost-to-income ratio was lower by 2007 relative to European peers.
- Regulatory capital to risk-weighted assets declined to just 7 percent at end-2007 and lagged behind peers.
- The banking system maintained a stable and positive return on assets, averaging 1 percent between 1999 and 2007.

### Funding and balance-sheet transformation
- Portuguese banks shifted from a retail-deposit model to a wholesale-funded model (interbank lending and bond issuance), enabled by euro area accession and cheaper euro-denominated wholesale funding.
- Loan-to-deposit ratio rose from 100 percent in 1999 to 150 percent in 2007, one of the highest in the euro area.

### Expansion of banking and misallocation of credit
- Banks’ total assets as a share of GDP rose from 184 percent of GDP in 1997 to 250 percent in 2007.
- Large capital inflows were intermediated mainly to domestic non-tradable sectors, contributing to:
  - Increase in the share of loans to real estate and construction from less than 13 percent of non-financial corporation (NFC) loans in 1997 to 38 percent in 2007.
- Loans to the construction sector increased from €8.4 billion in 1999 to €26.1 billion in 2008, while the sector’s value added stopped growing in the early 2000s and remained stable throughout the decade.

### Ownership structure and lending behavior
- The sector was dominated by five banking groups: Caixa Geral de Depósitos (CGD), Banco Comercial Português (BCP), Banco Espírito Santo (BES), Banco Português de Investimento (BPI) and Santander-Totta.
- Large economic groups were major shareholders, likely facilitating relationship-based lending to construction and real estate, consistent with findings that increases in domestic ownership can increase lending to less productive, non-tradable industries.

### Vulnerabilities entering the crisis
- Low capital levels and reliance on potentially volatile wholesale funding limited banks’ shock-absorption capacity.
- Credit concentrated in real estate and construction undermined asset quality and led to rising non-performing loans (NPLs).

### Crisis episode (2008–2013): crystallization of risks
- The global financial crisis caused a sudden stop of capital flows, economic slowdown, and higher funding costs, testing banking resilience.
- Reliance on wholesale funding resulted in liquidity issues and increased ECB funding utilization.
- Corporate sector excessive leverage led to loan defaults and a rise in NPLs, particularly in construction.
- Banks deleveraged (decline in outstanding loans) and sought deposits to reorient funding towards a retail base.
- NPLs to total loans nearly tripled from less than 4 percent in late-2008 to 11 percent by 2013, causing large impairment charges and declining provisioning coverage ratios.
- Net interest margins declined from 1.9 percent in 2008 to just 1.1 percent in 2013.
- The banking system experienced losses each year between 2011 and 2014.

### Empirical findings on drivers of profitability (2010–15)
- Data on 183 euro area banks (≥ €20 billion assets) show Portuguese banks had:
  - Lower capital ratios than most euro area banks (except Greece).
  - A higher share of deposits.
  - Higher RWA density due to lower asset quality and a more retail-oriented business.
  - Lower net income than most peers (except Greece, Spain and Ireland).
  - The highest operating costs to assets ratio (with Italy and Greece).
- Econometric results:
  - Bank-specific: size and current level of capital have the most positive impact on profitability; liquidity (deposits to assets) has a more limited contribution.
  - Macroeconomic: real GDP growth has a significant and large positive effect; public debt has a negative effect.
  - In stressed economies (Greece, Ireland, Italy, Portugal, Spain):
    - Capital’s impact on profitability is slightly larger.
    - Trading operations tend to reduce profitability (high exposure to domestic sovereign bonds).
    - Bank size is not significant.
    - Domestic public debt has a larger negative effect and EONIA has a large positive impact on profitability.
  - In non-stressed economies:
    - Bank size and capital are main factors; trading contributes positively.
    - Macroeconomic variables are generally not significant except VIX, which is positive.

### Public support and ownership changes
- Public support during the Economic Assistance Program (EAP) included the Bank Solvency Support Facility (BSSF) and private issuance.
- Public sector injected €7.2 billion into the banking system, mainly through contingent convertible debt instruments (€4.3 billion purchased by the BSSF).
- Private capital injections altered ownership:
  - BPI effectively controlled by CaixaBank and Angolan investors (19.5 percent for Santoro Finance).
  - BCP controlled by Sonangol (19.4 percent).
- Exposures to Angola increased significantly while exposures to Brazil declined markedly.

### Post-crisis (2014–16): constrained recovery
- Operating environment remained challenging despite moderate economic recovery.
- Banks reduced costs, but not sufficiently to offset declines in net interest income from low interest rates.
- Legacy issues persisted:
  - Debt of non-financial corporations rose to over 120 percent of GDP during the crisis and remained at 109 percent of GDP at end-2015.
  - NPLs continued to rise due to highly indebted non-financial corporations.
  - Bank capital levels remained among the lowest in the EU, limiting buffers to absorb losses.
- Bank resolutions:
  - BES resolved in August 2014 and Banif in December 2015.
  - Fiscal costs exceeded 4 percent of GDP: €4.9 billion for BES and €2.2 billion for Banif.
  - Contingent litigation and asset transfers to the Resolution Fund could entail further costs for government and banks.

### Medium-term challenges and policy implications
- Near-term expectations: economic growth, inflation and interest rates expected to remain low, increasing the need to strengthen the banking system to support recovery.
- Feedback loop: low profitability → difficulty increasing capital → slower balance sheet clean-up → continued vulnerability.
- Policy and bank actions recommended:
  - Bolster new and stable income streams:
    - Strengthen fee and commission income leveraging the universal banking model.
    - Capital Market Union and securitization markets could help develop fee-based business models, reduce capital requirements and improve risk-sharing.
  - Further reduce operating costs:
    - Compensation of bank employees accounts for a large share of revenue compared to other euro area countries.
    - Restructuring measures: rationalization of branch networks, divestment of non-core assets and unprofitable business lines.

*Source: PORTUGAL — INTERNATIONAL MONETARY FUND (content unit: _cr16301 - 2.       The drivers of bank profitability).*

### 18.      Banks will also need additional capital to absorb the losses stemming from changing

### 18.      Banks will also need additional capital to absorb the losses stemming from changing

### Banks’ capital needs and business-model adjustment
- Streamlining bank operations around core, profitable income streams and dealing with legacy assets will require significant loss absorption capacity by some banks.
- In the current environment, banks’ efforts to meaningfully bolster their capital buffers are hampered by the high cost of capital – as both new and existing shareholders are weary of providing additional capital to an unprofitable banking system with unresolved balance sheet weaknesses.
- Banks will need to move aggressively to restore profitability by identifying alternative stable income streams and lowering costs, while seeking a solution to long-standing legacy issues stemming from the misallocation of credit.
- Capital levels will need to be strengthened, likely even surpassing regulatory requirements, to absorb losses.

### Regulatory environment and short- versus medium-term effects
- Entry into force of the Bank Recovery and Resolution Directive, the Single Resolution Mechanism and future Minimum Requirements for own funds and Eligible Liabilities (MREL) impose additional requirements on banks.
- Short-run effect: regulatory requirements will impose costs on banks as they seek to adjust business models and shore up capital bases.
- Medium-term effect: regulatory requirements to further impair NPLs and raise additional capital would improve asset quality and free resources to lend to more productive firms, especially in the tradable sector.

### Context: Portugal and the ECB’s Asset Purchase Program (APP) / Quantitative Easing (QE)
- APP announced September 2014; initially monthly purchases of €60 billion to end-September 2016, extended to March 2017, with monthly purchases increasing to €80 billion from April 2016.
- APP coverage expanded from asset-backed securities and covered bonds to include public sector securities under PSPP (announced January 2015, began March 2015) and later corporate bonds under CSPP (launched June 2016).
- Average monthly purchases of Portuguese government bonds amounted to €1.2 billion from March 2015–June 2016.
- Total ECB holdings of Portuguese government bonds reached €19.1 billion at end-June 2016, equal to 17 percent of the total outstanding, with a weighted average remaining maturity of 10 years.
- Comparison: €19.1 billion (ECB holdings) versus €26.4 billion in IMF funding provided to Portugal under its 3-year adjustment program.
- Banks issued €1.8 billion in covered bonds and €0.7 billion in asset-backed securities since late 2014 programs began.
- Long-term loans to Portuguese banks provided under LTRO and TLTRO facilities amounted to €22 billion as of June 30.

### Transmission channels and anticipated effects of APP/QE
- Portfolio balance and credit channels: decline in sovereign bond yields expected to incentivize banks to shift from government bonds to riskier investments and ease lending conditions.
- Wealth channel: higher prices of financial and real estate assets expected to increase financial wealth, reducing precautionary savings and raising consumption.
- Confidence channel: ECB’s commitment expected to reduce uncertainty and encourage spending and investment.
- Exchange rate channel: lower interest rates expected to weaken the euro, supporting exports and inflation.

### Fiscal sector impacts observed
- Portugal’s net bond issuance: after limited issuance in 2013 (-€1.1 billion) and 2014 (€0.5 billion), net bond issuance jumped to €12.9 billion in the year prior to June 2016, almost fully covered by ECB purchases.
- Increased bond issuance allowed Portugal to make €10.3 billion in advanced repurchases to the Fund during 2015-16, around a third of its outstanding liability to the Fund at the end of the program, at an annual interest rate savings of around 0.1 percent of GDP through 2019.
- Ten-year Portuguese yields fell by 300 basis points during 2014; spreads over German bunds decreased by 200 basis points.
- Yields broadly stable at close to 3 percent throughout 2015-16.
- Juvenal and Wiseman (2015) estimate Portugal’s 10-year sovereign spread against German Bunds was 270 basis points below the level implied by macro fundamentals at end-March 2015.
- Bank of Portugal estimated that 2 and 10-year sovereign yields were about 2.5 percentage points below the level implied by macroeconomic fundamentals in October 2015.
- Fiscal adjustment: staff projections at 8th/9th review anticipated structural primary adjustment of 1.3 percent of GDP during 2015-16; actual path diverged to a structural primary easing of 1 percent of GDP projected for 2015-16.

### Financial sector impacts and constraints
- Interest rates on new loans to non-financial corporations of less than €1 million and new mortgage loans declined by 116 and 101 basis points, respectively, from December 2014–May 2016.
- New loans to households rose 30 percent (y/y) during the first half of 2016 after an increase of 28 percent (y/y) in 2015; growth concentrated in mortgage borrowing.
- New lending to NFCs decreased by 5 percent in the first half of 2015 and by 18 percent in 2015.
- NPLs to the non-financial corporate sector: 19.8 percent at end-2015, up from 18.6 percent at end-2014 and 17.2 percent at end-2013.
- NPLs to the household sector: relatively stable over the past four years at 7-8 percent; mortgage loan performance significantly better.
- Banks’ share of assets held in government securities has remained steady; little indication thus far of significant rebalancing in bank portfolios on aggregate.
- Decline in cost of funding (average rate on deposits, interbank and central bank financing) helped offset decrease in interest income, contributing to a small increase in banks’ net interest income in 2014-15 after large declines in 2012-2013.
- Risks ahead: limited space to further reduce deposit rates could begin to diminish bank profitability; a 100 basis point fall in Euribor rates estimated by Bank of Portugal to have a negative cumulative impact on domestic banks’ net interest income of approximately €700 million.
- Weaker profitability would limit scope for capital-constrained banks to be more proactive in NPL resolution.

### Key policy implications and required actions
- Banks must bolster loss-absorption capacity to manage legacy assets and business-model changes.
- Strengthen capital levels, likely beyond regulatory minima, to absorb losses.
- Implement aggressive cost reductions and identify alternative stable income streams.
- Regulatory reforms (BRRD, SRM, MREL) will impose short-term costs but support medium-term asset-quality improvement and reallocation of credit to more productive firms.
- Addressing high NPLs, especially in the corporate sector, is critical to unlock transmission of monetary easing to corporate lending and investment.

*Source: Excerpt from IMF staff report chapter on Portugal (text unit _cr16301 - 18).*

### 13.      QE appears to have had the strongest impact on consumption in Portugal. While

### _cr16301 - 13.      QE appears to have had the strongest impact on consumption in Portugal. While

### Impact on consumption and GDP composition
- Headline growth has not deviated from staff’s pre-QE projections, but the composition of growth became much more skewed towards consumption.
- The pickup in consumption appears to have mostly resulted in higher imports rather than stronger growth, reflecting the openness of Portugal’s economy and the high import content of consumption expenditure.
- Impact on investment appears relatively modest, relative to staff’s pre-QE projections, reflecting both the corporate debt overhang and the continued contraction in bank lending to NFCs.

- Portugal: Contributions to year-on-year growth, 2014-15 (percentage points, unless indicated otherwise)
  - Total domestic demand: 0.5 4.6 4.1
  - Final consumption expenditure: -0.3 3.2 3.5
  - Gross capital formation: 0.8 1.4 0.6
  - Foreign balance: 1.7 -2.3 -4.0
  - Real GDP growth, percent: 2.3 2.4 0.1

### Confidence and transmission channels
- QE (ECB APP and earlier signals including OMT) appears to have contributed to a significant recovery in consumer confidence, helping to allay concerns about the future of the area.
- Consumer confidence fell dramatically from early 2009 through mid-2012 as fears of a Eurozone exit rose, coinciding with a sharp fall in private consumption.
- ECB indications from mid-2012 and the launch of APP helped alleviate fears of an imminent break-up of the euro area, contributing to a sharp recovery in household consumption.

### Real estate, wealth effects, and household behavior
- Despite only a modest increase in disposable income of 1 percent between 2013Q1 and 2015Q3, household consumption grew by 6 percent.
- The corresponding decline in saving can be explained in part by the wealth effect from rising house prices: a one-percent increase in house prices corresponds to a roughly ½-percentage-point decline in the saving rate.
- The real estate wealth effect appears particularly important in Portugal given the higher real estate ownership rate.
- Recovery in housing prices since mid-2014 is attributable in part to recovery in confidence and employment, and record-low interest rates with a subsequent pickup in mortgage lending.

- Footnote data on real estate and financial asset effects:
  - Portugal has 5.9 million housing units with an estimated value of €300 billion.
  - House prices rose 9.9 percent between 2013Q1 and 2015Q4.
  - Accordingly, the estimated value of the real estate assets owned by households increased by €25 billion, while household savings declined by €5 billion.
  - On the financial asset side, estimated loss of €0.4 billion for a 50-bps decline in interest rates and a loss of €0.6 billion for a 100-bps decline.

### Credit channel, investment, and banking sector deleveraging
- The APP’s impact via the credit channel has been constrained by the corporate debt overhang.
- Transmission of lower interest rates to lending activity has been primarily limited to household mortgages and consumption loans.
- High corporate debt and rising NPLs on corporate loans have constrained lending to NFCs; net lending has continued to contract as increased availability of financing has done little to arrest deleveraging in the banking system.
- Investment has remained subdued despite the pick-up in consumption and decline in household savings.

### Macro impact, persistence, and limits
- The APP looks to have had a positive impact on consumer confidence and mortgage lending, boosting consumption through confidence and wealth channels.
- The impact on headline growth has been modest and is expected to fade as the household savings rate normalizes from historic lows at present.
- The impact on investment has been limited, reflecting the corporate debt overhang and rising uncertainty from the standstill in efforts to address structural obstacles to growth.

### Fiscal side-effects and policy implications
- Sovereign debt purchases through the PSPP will account for a large share of Portugal’s gross bond issuance in 2015-16, providing a backstop that alleviated financing concerns.
- This backstop has reduced the priority attached to fiscal adjustment: the structural primary adjustment in 2015-16 is now projected to be nearly 2½ percent of GDP less than originally envisaged prior to QE.
- Quoted policy-relevant observation: “For investment you need confidence, and for confidence you need structural reforms. The ECB has taken a further, very expansionary measure today, but it’s now up to the governments to implement these structural reforms, and the more they do, the more effective will be our monetary policy.” (ECB President Draghi, press conference announcing PSPP, January 2015)

*Source: IMF staff analysis in the provided chapter excerpt.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16301.pdf_