## 8. Impact of Financial Shocks by Household Wealth Decile

## Source details

**Canonical URL:** [8. Impact of Financial Shocks by Household Wealth Decile](https://www.imf.org/-/media/files/publications/wp/2018/wp18196.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2018/wp18196.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2018/wp18196.pdf.json)

---

### Introduction and context
- Total net household wealth at end-2013: over €9 trillion, or 5½ times GDP.
- Average wealth per household: exceeds €350,000; per capita: about €150,000.
- Asset composition:
  - Real assets (principally dwellings): almost two thirds of total net wealth.
  - Financial assets: remainder, mostly cash and deposits, shares, and insurance reserves.
- Distributional concentration:
  - Top two household wealth deciles hold more than two thirds of financial wealth and an even larger proportion of equity and non-secured debt.
  - Less wealthy households hold almost all financial wealth as insured bank deposits.
- Cross-sector context: high household net wealth contrasts with large negative net financial positions in other sectors (especially government and non-financial corporations), implying potential vulnerabilities via inter-sectoral linkages.

### Data, accounting method, and simulation approach
- Data sources and coverage:
  - Flow-of-funds: Bank of Italy quarterly series covering 1995:Q1–2017:Q3 (ESA 2010 classification).
  - Household distribution: Bank of Italy Survey on Household Income and Wealth (2016), representative sample ≈7,000 households; cross-country comparisons use ECB HFCS (2014).
- Sectors: NFC, MFI, BOI, OFI, INP, GOV, HH, RoW.
- Exposure matrix: entry (i,j) = assets of sector i invested in sector j (liabilities of j to i).
- Simulation mechanics:
  - Direct NFA impact = holdings of shocked instrument; indirect impact = propagation via equity linkages.
  - Ultimate equity owners (absorbers): HH, GOV, BOI, RoW; non-absorbers: NFC, MFI, OFI, INP.
  - Bank bail-in scenario: MFI equity wiped out; bonds converted to equity at conversion rate of 50 percent; no public resolution funds; indirect equity impact for MFIs = 0.
  - Bank bailout scenario: government compensates MFI bondholders for losses equal to those in bail-in; shareholders wiped out; indirect equity effect for MFIs is negative (minus the government transfer).
  - Simulations abstract from correlated price changes (beyond equity), portfolio reallocations, detailed bank-level contagion, and real-economy feedbacks.

### Sectoral financial exposures (2017:Q3) — key statistics
- Aggregate total financial assets (sum across sectors): 843 percent of GDP.
- Selected sector NFA positions (percent of GDP):
  - HH NFA: 196
  - GOV NFA: -124
  - NFC NFA: -112
  - RoW NFA: 8
  - BOI NFA: 6
  - INP NFA: -2
  - MFI NFA: 0
- Representative exposure entries (percent of GDP, 2017:Q3):
  - HH holdings of MFI: 64
  - HH holdings of NFC: 54
  - RoW holdings of GOV: 47
  - MFI liabilities to RoW: 50
  - Total liabilities (system): 843 percent of GDP
- Portfolio composition (percent of sector assets, 2017:Q3):
  - MFIs: 22 percent of assets in GOV.
  - INP: 36 percent of assets in GOV.
  - HH portfolio: 26 percent in MFI, 20 percent in INP, 22 percent in NFC.

### Evolution of exposures (selected snapshots)
- Snapshots: 1995:Q1; 2001:Q4; 2008:Q1; 2012:Q2; 2017:Q3.
- Total financial assets growth: from 418 to 843 percent of GDP (1995:Q1 → 2017:Q3).
- Key trends:
  - Growing importance of foreign liabilities and foreign asset holdings; international diversification peaked around Euro accession but slowed thereafter (home bias remains elevated).
  - Households shifted assets away from GOV into INP; MFI holdings peaked pre-GFC and receded thereafter.
  - GOV funding increasingly reliant on MFIs post-GFC; since 2012 BOI became a major GOV creditor reflecting QE.

### Illustrative shock scenarios (2017:Q3 data) — quantified impacts
- Government bond value shock: 10 percent decline in general government bond value (equivalent to yield increase ≈220 basis points given duration 4.88 years).
  - Impacts (percent of GDP; direct, equity, ∆NFA = direct + equity):
    - NFC: direct -0.3; equity 0.3; ∆NFA 0.0
    - MFI: direct -2.1; equity 2.1; ∆NFA 0.0
    - BOI: direct -2.2; equity 0.0; ∆NFA -2.2
    - OFI: direct -0.7; equity 0.7; ∆NFA 0.0
    - INP: direct -2.0; equity 2.0; ∆NFA 0.0
    - GOV: direct 11.7; equity -0.4; ∆NFA 11.3
    - HH: direct -0.8; equity -3.1; ∆NFA -3.8
    - RoW: direct -3.7; equity -1.6; ∆NFA -5.3
  - Interpretation:
    - Direct losses concentrated in financial sector and RoW; equity linkages shift larger total NFA decline to RoW and HH.
    - Household impact ≈4 percent of GDP (≈2 percent of household NFA).

- Bank bail-in: banks constituting 10 percent of MFI assets are bailed in (equity wiped out; bonds converted to equity at 50 percent).
  - Impacts (percent of GDP):
    - NFC: direct -0.2; equity 0.2; ∆NFA 0.0
    - MFI: direct 2.3; equity 0.0; ∆NFA 2.3
    - BOI: direct -0.1; equity 0.0; ∆NFA -0.1
    - OFI: direct -0.1; equity 0.1; ∆NFA 0.0
    - INP: direct -0.1; equity 0.1; ∆NFA 0.0
    - GOV: direct -0.1; equity 0.0; ∆NFA -0.1
    - HH: direct -0.9; equity -0.3; ∆NFA -1.2
    - RoW: direct -0.7; equity -0.1; ∆NFA -0.9
  - Interpretation:
    - Bail-in burden mainly absorbed by HH and RoW; almost half of bail-in’s impact absorbed by RoW, mitigating domestic real-economy damage.
    - HH impact ≈1.2 percent of GDP (≈0.5 percent of household NFA).
    - MFI NFA rises because liabilities to other sectors are reduced.

- Bank bailout: banks constituting 10 percent of MFI assets are bailed out (equity wiped out; government compensates bondholders for 50 percent of bond value).
  - Impacts (percent of GDP):
    - NFC: direct -0.2; equity 0.2; ∆NFA 0.0
    - MFI: direct 2.9; equity -0.6; ∆NFA 2.3
    - BOI: direct 0.0; equity 0.0; ∆NFA 0.0
    - OFI: direct 0.0; equity 0.0; ∆NFA 0.0
    - INP: direct 0.0; equity 0.0; ∆NFA 0.0
    - GOV: direct -1.7; equity 0.0; ∆NFA -1.6
    - HH: direct -0.6; equity 0.2; ∆NFA -0.4
    - RoW: direct -0.3; equity 0.1; ∆NFA -0.2
  - Interpretation:
    - Bailout shifts most burden to GOV (over 1½ percent of GDP) and domestic HH.
    - Foreign absorption falls by 75 percent compared to bail-in, as government transfers also go to foreign bondholders.

- Effect of recapitalization: 50 percent increase in aggregate capital-to-assets ratio of MFIs
  - Cuts the public cost of a bailout by a third.
  - Doubles the foreign absorption rate.
  - Slightly increases the impact of a bail-in on domestic households (households hold a disproportionate amount of equity relative to bonds compared to foreigners).
  - Recapitalization could also reduce banks’ risk-taking and thereby the probability/size of shocks (outside the static accounting exercise).

### Bank shock counterfactuals (bond conversion rate scenarios)
- Appendix IV compares baseline conversion rate (50 percent) with zero conversion rate ("no convers.").
- A lower conversion rate increases total bondholder losses and raises required government transfer in a bailout.
- Bail-in, baseline vs no convers. (percent of GDP):
  - NFC: 0.0 vs 0.0
  - MFI: 2.3 vs 3.2
  - BOI: -0.1 vs -0.2
  - OFI: 0.0 vs 0.0
  - INP: 0.0 vs 0.0
  - GOV: -0.1 vs -0.1
  - HH: -1.2 vs -1.6
  - RoW: -0.9 vs -1.3
- Bailout, baseline vs no convers. (percent of GDP):
  - NFC: 0.0 vs 0.0
  - MFI: 2.3 vs 3.2
  - BOI: 0.0 vs 0.0
  - OFI: 0.0 vs 0.0
  - INP: 0.0 vs 0.0
  - GOV: -1.6 vs -3.2
  - HH: -0.4 vs 0.0
  - RoW: -0.2 vs 0.0

### Distributional impacts across households and deciles
- Concentration of financial instruments (top 20 percent):
  - Hold 69 percent of financial wealth.
  - Hold 93 percent of equity.
  - Hold 79 percent of bank bonds.
  - Government bonds distributed in line with total financial wealth.
- Portfolio patterns by wealth:
  - Less-wealthy households: portfolio almost entirely insured bank deposits (limited direct exposure to GOV or bank shocks).
  - Wealthier households: invest proportionally more in equity and non-secured fixed income; most financial wealth still in safe assets.
- Consequence for shock absorption:
  - Wealthy households absorb most financial losses after a government bond shock or a bail-in.
  - Social welfare effects are smaller than wealth effects because losses concentrate at the top.
  - Cost of a bailout is less concentrated among the wealthiest; top decile contributes less as a percentage of their financial assets than the next three deciles due to government compensation channels and positive indirect equity impacts.

- Table 8 — impacts by household wealth decile (per-household euros; % of decile financial assets)
  - Decile 1: Wealth -1,584; Fin. As. 477; Gov. bond 0 euros; Bail-in 0 euros; Bailout 0 euros; Gov. bond -0.1%; Bail-in 0.0%; Bailout 0.0%
  - Decile 2: Wealth 3,326; Fin. As. 2,438; Gov. bond 0 euros; Bail-in 0 euros; Bailout 0 euros; Gov. bond 0.0%; Bail-in 0.0%; Bailout 0.0%
  - Decile 3: Wealth 17,651; Fin. As. 9,488; Gov. bond -51 euros; Bail-in -23 euros; Bailout -12 euros; Gov. bond -0.5%; Bail-in -0.2%; Bailout -0.1%
  - Decile 4: Wealth 65,334; Fin. As. 8,486; Gov. bond -76 euros; Bail-in -18 euros; Bailout -8 euros; Gov. bond -0.9%; Bail-in -0.2%; Bailout -0.1%
  - Decile 5: Wealth 108,434; Fin. As. 9,928; Gov. bond -91 euros; Bail-in -61 euros; Bailout -36 euros; Gov. bond -0.9%; Bail-in -0.6%; Bailout -0.4%
  - Decile 6: Wealth 146,859; Fin. As. 13,849; Gov. bond -183 euros; Bail-in -92 euros; Bailout -46 euros; Gov. bond -1.3%; Bail-in -0.7%; Bailout -0.3%
  - Decile 7: Wealth 194,062; Fin. As. 17,044; Gov. bond -306 euros; Bail-in -126 euros; Bailout -62 euros; Gov. bond -1.8%; Bail-in -0.7%; Bailout -0.4%
  - Decile 8: Wealth 254,080; Fin. As. 24,771; Gov. bond -504 euros; Bail-in -277 euros; Bailout -149 euros; Gov. bond -2.0%; Bail-in -1.1%; Bailout -0.6%
  - Decile 9: Wealth 367,042; Fin. As. 45,014; Gov. bond -1,260 euros; Bail-in -650 euros; Bailout -310 euros; Gov. bond -2.8%; Bail-in -1.4%; Bailout -0.7%
  - Decile 10: Wealth 914,511; Fin. As. 146,949; Gov. bond -8,183 euros; Bail-in -2,113 euros; Bailout -511 euros; Gov. bond -5.6%; Bail-in -1.4%; Bailout -0.3%
  - Total (per household averages): Wealth 206,971; Fin. As. 27,844; Gov. bond -1,066 euros; Bail-in -336 euros; Bailout -113 euros; Gov. bond -3.8%; Bail-in -1.2%; Bailout -0.4%
- Data notes:
  - First column: household total net wealth (including real assets).
  - Second column: gross financial assets.
  - Results assume bank equity exposures distributed as total equity exposures.
  - Survey-level financial asset levels are lower than 2013 Bank of Italy aggregates; survey used only for distribution, flow-of-funds used for financial wealth levels.

### Key substantive findings and policy implications
- Household capacity:
  - Households, especially at the upper end, have ample capacity to absorb plausible financial shocks given large HH NFA and concentration of financial wealth among wealthy households.
- Shock-specific summaries:
  - A 10 percent fall in government bond values produces household ∆NFA of -3.8 percent of GDP (≈2 percent of household NFA); RoW absorbs -5.3 percent of GDP.
  - Bail-ins share shock burdens more with RoW and reduce domestic public costs relative to bailouts.
  - Bailouts concentrate costs on GOV (public debt) and transfer resources to wealthier households and to RoW.
- Fiscal and distributional considerations:
  - Government bailouts are fiscally costly and regressive: adding to public debt imposes costs via taxes on labor income (which fall more evenly across the distribution than wealth effects) and has weak countercyclical implications because wealthier households have a lower marginal propensity to consume.
- Role of recapitalization:
  - Recapitalizing banks can materially reduce public bailout costs and raise foreign absorption, though it may modestly increase household exposure in bail-in scenarios.
  - Recapitalization may also lower the probability and size of future shocks by limiting bank risk-taking (not captured in static accounting).
- Caveats:
  - The balance-sheet accounting exercise abstracts from real effects, correlated asset price movements (beyond equity), portfolio reallocation, contagion across individual banks, and dynamics—policy choices (bail-ins vs. bailouts) must consider all channels and financial stability tradeoffs.
  - Bail-in and bailout scenarios are illustrative polar cases; actual restructurings typically contain elements of both.
  - Smaller banks, with more vulnerable balance sheets and disproportionate domestic household ownership, would reduce the subsidy to RoW associated with a bailout if modeled explicitly.

*Italic: IMF Working Paper — "8. Impact of Financial Shocks by Household Wealth Decile" (excerpted content unit).*

### REFERENCES ______________________________________________________________19

### REFERENCES ______________________________________________________________19

### Figures
- 1. Average Net Household Wealth by Decile, 2014 _________________________________4
- 2. Net Financial Assets by Sector _______________________________________________5
- 3. Select Sectoral Financial Exposures __________________________________________10
- 4. Distribution of Household Assets by Household Wealth, 2016:Q4 __________________16
- 5. Household Financial Asset Portfolio by Wealth, 2016:Q4 _________________________16

### Tables
- 1. Grouping of Sectors _______________________________________________________7
- 2. Sectoral Financial asset Exposures, 2017:Q3 ____________________________________8
- 3. Sectoral Financial Asset Exposures, 2017:Q3 ___________________________________9
- 4. Impact of a Government Bond Value Shock ___________________________________12
- 5. Impact of a Bank Bail-in ___________________________________________________13
- 6. Impact of a Bank Bailout __________________________________________________13
- 7. Impact of Bank Shocks, Recapitalized Banks Counterfactual ______________________15

*Source: wp18196 - REFERENCES ______________________________________________________________19*

### 8. Impact of Financial Shocks by Household Wealth Decile _________________________17

### 8. Impact of Financial Shocks by Household Wealth Decile

### Introduction and context
- Total net household wealth at end-2013: over €9 trillion, or 5½ times GDP.
- Average wealth per household: exceeds €350,000; per capita: about €150,000.
- Real assets (principally dwellings) constitute almost two thirds of total net wealth; remainder are financial assets (mostly cash and deposits, shares, and insurance reserves).
- The top two household wealth deciles hold more than two thirds of financial wealth and an even larger proportion of equity and non-secured debt; less wealthy households hold almost all financial wealth as insured bank deposits.
- High household net wealth contrasts with large negative net financial positions in other sectors (especially government and non-financial corporations), implying potential financial vulnerabilities through inter-sectoral linkages.

### Data and accounting method
- Flow-of-funds data: Bank of Italy quarterly series covering 1995:Q1–2017:Q3 (European System of Accounts 2010 classification).
- Sectors used include: NFC, MFI, BOI, OFI, INP, GOV, HH, RoW.
- Constructed matrix of cross-sectoral bilateral financial exposures where entry (i,j) = assets of sector i invested in sector j (equivalently liabilities of j to i).
- Household distributional analysis uses Bank of Italy Survey on Household Income and Wealth (2016), representative sample ≈7,000 households; cross-country comparisons use ECB Household Finance and Consumption Survey (2014).
- Simulations compute direct, indirect (equity-linkage) and total impacts on sectoral NFA; ultimate equity owners are HH, GOV, BOI and RoW, which absorb the total NFA impact.

### Sectoral financial exposures (2017:Q3) — key statistics
- Aggregate total financial assets (sum across sectors): 843 percent of GDP.
- Selected sector net financial asset (NFA) positions (percent of GDP):
  - HH NFA: 196
  - GOV NFA: -124
  - NFC NFA: -112
  - RoW NFA: 8
  - BOI NFA: 6
  - INP NFA: -2
  - MFI NFA: 0
- Representative rows/columns from the sectoral exposure matrix (percent of GDP, 2017:Q3):
  - HH holdings of MFI: 64
  - HH holdings of NFC: 54
  - RoW holdings of GOV: 47
  - MFI liabilities to RoW: 50
  - Total liabilities (system): 843 percent of GDP
- Portfolio composition (percent of sector assets, 2017:Q3) highlights:
  - MFIs: 22 percent of assets in GOV (government exposure measured as share of MFI assets).
  - INP (insurance and pension): 36 percent of assets in GOV.
  - HH portfolio: 26 percent in MFI, 20 percent in INP, 22 percent in NFC.

### Evolution of exposures (selected snapshots)
- Five snapshots: 1995:Q1; 2001:Q4; 2008:Q1; 2012:Q2; 2017:Q3.
- Total financial assets doubled from 418 to 843 percent of GDP (1995:Q1 → 2017:Q3).
- Key trends:
  - Growing importance of foreign liabilities and foreign asset holdings; international diversification peaked around Euro accession but slowed thereafter (home bias remains elevated).
  - Households shifted assets away from GOV and into INP; MFI holdings peaked pre-GFC and receded thereafter.
  - GOV funding increasingly reliant on MFIs post-GFC; since 2012 BOI became a major GOV creditor reflecting QE.

### Simulation method for shocks
- Valuation shock to instruments issued by a sector: direct NFA impact = holdings of that instrument; indirect impact = changes via inter-sectoral equity linkages.
- For non-ultimate equity sectors, final NFA impact is zero once equity linkages propagate to ultimate shareholders (HH, GOV, BOI, RoW).
- Bank bail-in scenario: affected MFIs’ equity wiped out; all their bonds converted into equity at conversion rate of 50 percent; no public resolution funds used; other (non-bond) debtholders not affected. Indirect equity impact for MFIs = 0.
- Bank bailout scenario: government transfers to compensate MFI bondholders for losses equal to those in bail-in; shareholders wiped out; indirect equity effect for MFIs is negative (minus the government transfer to troubled-bank bondholder MFIs).
- Simulations abstract from correlated price changes (beyond equity), portfolio reallocations, contagion across individual banks, and dynamic real-economy feedbacks.

### Illustrative shock scenarios and quantified impacts (2017:Q3 data)
- Government bond value shock: assume a 10 percent decline in the value of general government bonds.
  - Rationale: sizable shock; equivalent to an increase in yields of around 220 basis points given current duration 4.88 years.
  - Table of impacts (percent of GDP) — direct impact, equity impact, ∆NFA (direct + equity):
    - NFC: direct -0.3; equity 0.3; ∆NFA 0.0
    - MFI: direct -2.1; equity 2.1; ∆NFA 0.0
    - BOI: direct -2.2; equity 0.0; ∆NFA -2.2
    - OFI: direct -0.7; equity 0.7; ∆NFA 0.0
    - INP: direct -2.0; equity 2.0; ∆NFA 0.0
    - GOV: direct 11.7; equity -0.4; ∆NFA 11.3
    - HH: direct -0.8; equity -3.1; ∆NFA -3.8
    - RoW: direct -3.7; equity -1.6; ∆NFA -5.3
  - Interpretation: direct losses concentrated in financial sector and RoW; accounting for equity linkages shifts larger total NFA decline to RoW and HH. Household impact ≈4 percent of GDP but ≈2 percent of household NFA.

- Bank bail-in: assume banks constituting 10 percent of MFI assets are bailed in (all equity wiped out; bonds converted to equity at 50 percent).
  - Table of impacts (percent of GDP):
    - NFC: direct -0.2; equity 0.2; ∆NFA 0.0
    - MFI: direct 2.3; equity 0.0; ∆NFA 2.3
    - BOI: direct -0.1; equity 0.0; ∆NFA -0.1
    - OFI: direct -0.1; equity 0.1; ∆NFA 0.0
    - INP: direct -0.1; equity 0.1; ∆NFA 0.0
    - GOV: direct -0.1; equity 0.0; ∆NFA -0.1
    - HH: direct -0.9; equity -0.3; ∆NFA -1.2
    - RoW: direct -0.7; equity -0.1; ∆NFA -0.9
  - Interpretation: bail-in impact mainly absorbed by HH and RoW; almost half of bail-in’s impact is absorbed by RoW, mitigating domestic real economy damage. HH impact ≈1.2 percent of GDP (≈0.5 percent of household NFA). MFI NFA rises because liabilities to other sectors are reduced.

- Bank bailout: assume banks constituting 10 percent of MFI assets are bailed out (equity wiped out; government compensates bondholders for 50 percent of bond holdings value).
  - Table of impacts (percent of GDP):
    - NFC: direct -0.2; equity 0.2; ∆NFA 0.0
    - MFI: direct 2.9; equity -0.6; ∆NFA 2.3
    - BOI: direct 0.0; equity 0.0; ∆NFA 0.0
    - OFI: direct 0.0; equity 0.0; ∆NFA 0.0
    - INP: direct 0.0; equity 0.0; ∆NFA 0.0
    - GOV: direct -1.7; equity 0.0; ∆NFA -1.6
    - HH: direct -0.6; equity 0.2; ∆NFA -0.4
    - RoW: direct -0.3; equity 0.1; ∆NFA -0.2
  - Interpretation: bailout shifts most burden to GOV (over 1½ percent of GDP) and domestic HH; foreign absorption falls by 75 percent compared to bail-in, as government transfers also go to foreign bondholders.

- Effect of recapitalization: a 50 percent increase in aggregate capital-to-assets ratio of MFIs
  - Cuts the public cost of a bailout by a third.
  - Doubles the foreign absorption rate.
  - Slightly increases the impact of a bail-in on domestic households (households hold a disproportionate amount of equity relative to bonds compared to foreigners).
  - Recapitalization could also reduce banks’ risk-taking and thereby the probability/size of shocks (outside the static accounting exercise).

### Key substantive findings and policy implications
- Households, especially at the upper end of the distribution, have ample capacity to absorb plausible financial shocks given the large household NFA and concentration of financial wealth among wealthy households.
- A 10 percent fall in government bond values produces a household ∆NFA of -3.8 percent of GDP (≈2 percent of household NFA); the rest of the world absorbs a large share (-5.3 percent of GDP).
- Bail-ins share shock burdens more with RoW and reduce domestic public costs relative to bailouts; bailouts concentrate costs on GOV (public debt) and transfer resources to wealthier households and to RoW.
- Government intervention via bailouts is fiscally costly and regressive: adding to public debt imposes costs via taxes on labor income (which fall more evenly across the distribution than wealth effects) and has weak countercyclical implications because wealthier households have a lower marginal propensity to consume.
- Recapitalizing banks can materially reduce public bailout costs and raise foreign absorption, though it may modestly increase household exposure in bail-in scenarios; recapitalization may also lower the probability and size of future shocks by limiting bank risk-taking.
- The simple balance-sheet accounting exercise abstracts from real effects, correlated asset price movements (beyond equity), portfolio reallocation, contagion across individual banks, and dynamics—policy decisions (bail-ins vs. bailouts) must consider all channels and financial stability tradeoffs.

*Italic source: IMF Working Paper — "8. Impact of Financial Shocks by Household Wealth Decile" (excerpted content unit).*

### Appendix IV presents results assuming a zero conversion rate. Second, smaller banks, which

### Appendix IV presents results assuming a zero conversion rate. Second, smaller banks, which

### Bank shock counterfactuals and bond conversion scenarios
- Appendix IV compares a baseline assumption (50 percent recovery of bond value after conversion to equity) with a zero conversion rate ("no convers.").
- A lower conversion rate increases total bondholder losses and raises the required government transfer in a bailout.
- In the bail-in case, the additional impact from a zero conversion rate is roughly proportionally distributed across sectors.

- Table A1 (Impact of Bank Shocks, No Bond Conversion Counterfactual) — Percent of GDP:
  - Bail-in, baseline vs no convers.:
    - NFC: 0.0 vs 0.0
    - MFI: 2.3 vs 3.2
    - BOI: -0.1 vs -0.2
    - OFI: 0.0 vs 0.0
    - INP: 0.0 vs 0.0
    - GOV: -0.1 vs -0.1
    - HH: -1.2 vs -1.6
    - RoW: -0.9 vs -1.3
  - Bailout, baseline vs no convers.:
    - NFC: 0.0 vs 0.0
    - MFI: 2.3 vs 3.2
    - BOI: 0.0 vs 0.0
    - OFI: 0.0 vs 0.0
    - INP: 0.0 vs 0.0
    - GOV: -1.6 vs -3.2
    - HH: -0.4 vs 0.0
    - RoW: -0.2 vs 0.0

### Distributional impact across sectors and households
- Households and sectoral exposures:
  - Households at the top 20 percent of the wealth distribution hold 69 percent of financial wealth.
  - Equity holdings concentrated at the top 20 percent: 93 percent.
  - Bank bonds concentrated at the top 20 percent: 79 percent.
  - Government bonds distributed in line with total financial wealth.

- Portfolio composition by wealth:
  - Less-wealthy households: financial portfolio almost entirely insured bank deposits (limited direct exposure to government bond or bank financial shocks).
  - Wealthier households: invest proportionally more in equity and non-secured fixed income instruments, though most of their financial wealth is also in safe assets.

- Consequence for shock absorption:
  - Wealthy households absorb most financial losses after a government bond shock or a bail-in.
  - Social welfare effects are smaller than wealth effects because losses concentrate at the top of the wealth distribution.
  - The cost of a bailout is less concentrated among the wealthiest households; top decile contributes less as a percentage of their financial assets than the next three deciles due to government compensation channels and positive indirect equity impacts.

- Table 8 (Impact of Financial Shocks by Household Wealth Decile) — percentage points (euros per HH; % of decile fin. assets):
  - Decile 1: Wealth -1,584; Fin. As. 477; Gov. bond 0 euros; Bail-in 0 euros; Bailout 0 euros; Gov. bond -0.1%; Bail-in 0.0%; Bailout 0.0%
  - Decile 2: Wealth 3,326; Fin. As. 2,438; Gov. bond 0 euros; Bail-in 0 euros; Bailout 0 euros; Gov. bond 0.0%; Bail-in 0.0%; Bailout 0.0%
  - Decile 3: Wealth 17,651; Fin. As. 9,488; Gov. bond -51 euros; Bail-in -23 euros; Bailout -12 euros; Gov. bond -0.5%; Bail-in -0.2%; Bailout -0.1%
  - Decile 4: Wealth 65,334; Fin. As. 8,486; Gov. bond -76 euros; Bail-in -18 euros; Bailout -8 euros; Gov. bond -0.9%; Bail-in -0.2%; Bailout -0.1%
  - Decile 5: Wealth 108,434; Fin. As. 9,928; Gov. bond -91 euros; Bail-in -61 euros; Bailout -36 euros; Gov. bond -0.9%; Bail-in -0.6%; Bailout -0.4%
  - Decile 6: Wealth 146,859; Fin. As. 13,849; Gov. bond -183 euros; Bail-in -92 euros; Bailout -46 euros; Gov. bond -1.3%; Bail-in -0.7%; Bailout -0.3%
  - Decile 7: Wealth 194,062; Fin. As. 17,044; Gov. bond -306 euros; Bail-in -126 euros; Bailout -62 euros; Gov. bond -1.8%; Bail-in -0.7%; Bailout -0.4%
  - Decile 8: Wealth 254,080; Fin. As. 24,771; Gov. bond -504 euros; Bail-in -277 euros; Bailout -149 euros; Gov. bond -2.0%; Bail-in -1.1%; Bailout -0.6%
  - Decile 9: Wealth 367,042; Fin. As. 45,014; Gov. bond -1,260 euros; Bail-in -650 euros; Bailout -310 euros; Gov. bond -2.8%; Bail-in -1.4%; Bailout -0.7%
  - Decile 10: Wealth 914,511; Fin. As. 146,949; Gov. bond -8,183 euros; Bail-in -2,113 euros; Bailout -511 euros; Gov. bond -5.6%; Bail-in -1.4%; Bailout -0.3%
  - Total (per household averages): Wealth 206,971; Fin. As. 27,844; Gov. bond -1,066 euros; Bail-in -336 euros; Bailout -113 euros; Gov. bond -3.8%; Bail-in -1.2%; Bailout -0.4%

- Notes on data:
  - First column in Table 8: household total net wealth (including real assets).
  - Second column: gross financial assets.
  - Results assume bank equity exposures are distributed as total equity exposures.
  - Survey-level financial asset levels are lower than 2013 Bank of Italy aggregates; survey used only for distribution, flow-of-funds data used for financial wealth levels.

### Comparison of bail-in and bailout effects and policy implications
- Bail-in:
  - Overall impact on wealth is relatively small.
  - Impact on welfare is even smaller, given incidence concentrated at the top end of the wealth distribution.
  - Degree of domestic absorption in a bail-in is much lower than in a bailout.

- Bailout:
  - Government transfers resources (via transfers to bondholders), effectively transferring resources to foreigners in the studied setup and thus exposing the government (and taxpayers) to larger losses.
  - Costs of a bailout will need to be passed on to the domestic taxpayer, spreading losses across broader parts of the income and wealth distribution given reliance on labor income and consumption taxes rather than wealth taxes.
  - Bailouts may prevent contagion to other MFIs, potentially averting detrimental knock-on effects on financial stability and investment—this is a consideration that distinguishes bailouts from bail-ins.

- Interpretation:
  - Bail-in and bailout scenarios are illustrative polar cases; actual bank restructuring typically contains elements of both.
  - Smaller banks, with more vulnerable balance sheets and disproportionate domestic household ownership, would reduce the subsidy to the rest of the world associated with a bailout if taken into account.

### Methodological highlights (shock identification and transmission)
- Valuation shock direct impact formula:
  - d_ijk = ∆p_jk * a_ijk, if i ≠ j
  - d_jjk = −∆p_jk * l_jk, for i = j
  - ∆p_jk is change in value of instrument k issued by sector j; a_ijk holdings of sector i in those liabilities; l_jk total instrument-k liabilities of sector j.
- Indirect impact through equity linkages:
  - Sectors divided into ultimate equity owners {BOI, GOV, HH, RoW} and non-absorbers {NFC, MFI, OFI, INP}.
  - Intermediate equity impact among non-absorbers: e' = (I − M)−1 ∑ d_sjk E'_{is}, with matrices M and E' defined as in the text to capture equity holdings and liabilities and infinite rounds of knock-on effects.
  - Ultimate equity impact for absorbers: e_i = ∑ e'_s E_{is}, with E imposing zero ownership for non-absorber sectors.
  - For non-absorbers, equity impact equals −d_ijk (change in assets mirrors equity liabilities).
- Bail-in specific assumptions:
  - MFI equity is wiped out; e_i = 0 for i = MFI.
  - Intermediate equity impact recalculated excluding MFIs as absorbers.
- Bailout specific assumptions:
  - Government transfer t equals bond value loss: t = ∆p_jk ∑_i a_ijk (j = MFI, k = bond).
  - Government direct impact is −t; other sectors’ direct impacts given by exposure to MFI equity.
  - Ultimate equity impacts adjusted to account for government transfer being transmitted to MFI shareholders.

_italic: Source: Bank of Italy (via Haver) and staff calculations, as presented in the provided IMF working paper content._

---


_Source: https://www.imf.org/-/media/files/publications/wp/2018/wp18196.pdf_