## wpiea2021085-print-pdf

## Source details

**Canonical URL:** [wpiea2021085-print-pdf](https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021085-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2021/english/wpiea2021085-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2021/english/wpiea2021085-print-pdf.pdf.json)

---

### I Introduction and research question
- Focus: whether a tax on parental wealth affects the labor income distribution of the next generation in Norway.
- Cohorts analyzed: born during 1978-1980.
- Timing of parental taxation exposure: parents taxed in the late 1990s (and early 2000s) when children were at advanced stages in school.
- Outcome observation window for children in adulthood: 2013-2017.
- Primary outcomes examined:
  - the level of wage;
  - the position on the labor income distribution;
  - position on the labor income distribution relative to that of the parents (intergenerational income mobility).

### Identification strategy and empirical design
- Sources of variation:
  - changes to the wealth tax rate in the late 1990s and early 2000s; and
  - different levels of taxation of the same level of wealth depending on marital status (married joint filers had higher threshold and deduction than single filers).
- Estimation approach:
  - IV model using tax changes (marital status differences and tax law changes) as an instrument for changes in net parental wealth.
  - Instrument notation: instrument Z_i is parental divorce (change in tax treatment due to marital status change).
  - Outcome Y_i: wages. Parental wealth X_i. Confounder C_i: unobservables affecting wages and related to parental wealth.
- Concern and correction for exclusion restriction:
  - If divorce directly affects children's earnings (δ≠0), IV biased.
  - Strategy: estimate δ from a sample with parental wealth below the tax threshold (divorce does not affect tax payments) via Y_j = α + β X_j + δ Z_j + e_j to identify δ̂.
  - Construct adjusted outcome Y_pred_i = Y_i − δ̂ Z_i and use Y_pred_i in IV second stage to identify β.
  - Authors note this provides an upper bound of the divorce effect if δ̂ is larger for the low-wealth sample; if δ constant across samples adjustment yields unbiased δ.
- Specifications and samples:
  - Cohorts: 1978–1980 (ages 14–16 in 1993, 19–21 in 1998, 38–40 in 2017).
  - OLS baseline controls include parental wages, education, age, marital status in 1993; individual controls include age and born outside Norway; year dummies included.
  - IV instrument constructed as ∆tax payment_p,i,t = taxrule_t(netw_p,i,1993) − taxrule_1993(netw_p,i,1993).
  - Conditioning on both parents alive in 1998 so variation in tax treatment is driven by divorce.
- Instrument diagnostics:
  - First-stage F−statistics and R2 support relevance; pass Stock-Yogo cutoffs (see Appendix Table A.5).

### Main empirical findings (summary of main tables and text)
- Overall interpretation: Using exogenous changes in parental net wealth, children from wealthier families tend to have higher labor income; a wealth tax reduces that intergenerational advantage.
- Key point estimates (monetary amounts in Table 1 measured in NOK 1000; prose provides USD conversions):
  - Effect of net parental wealth on wage level (total net wealth):
    - OLS: 0.00271*** (0.000356)
    - IV: 0.0167*** (0.00121)
    - Adjusted IV: 0.0140*** (0.00119)
    - Prose USD interpretations: OLS implies USD 27,100; IV USD 16,700; Adjusted IV USD 14,000 (per USD 1 million net parental wealth).
  - Effect of parental financial wealth on wage level:
    - OLS: 0.00255*** (0.000395)
    - IV: 0.0179*** (0.00124)
    - Adjusted IV: 0.0101*** (0.00103)
    - Prose USD interpretations: USD 25,500 (OLS), USD 17,900 (IV), USD 10,100 (Adjusted IV) per USD 1 million.
  - Effect on wage percentile:
    - Net parental wealth: OLS 0.000208*** (0.0000276), IV 0.00135*** (0.000104), Adjusted IV 0.00113*** (0.000107).
    - Parental financial wealth: OLS 0.000218*** (0.0000333), IV 0.00145*** (0.000105), Adjusted IV 0.000745*** (0.000890).
  - Effect on wage mobility:
    - Net parental wealth: OLS 0.000649*** (0.0000638), IV 0.00317*** (0.000180), Adjusted IV 0.00295*** (0.000191).
    - Parental financial wealth: OLS 0.000633*** (0.0000869), IV 0.00335*** (0.000190), Adjusted IV 0.00206*** (0.000158).
- Sample sizes in Table 1:
  - N = 480,971 (columns 1,2,4,5); N = 270,995 (columns 3,6).
- Inequality counterfactual:
  - Observed Gini coefficient close 0.24; counterfactual Gini coefficient in the sample close to 0.25 (based on adjusted IV, column 3).
- Controls: parental wages, education and age of father and mother, parental wealth and marital status in 1993, whether the individual earns mainly capital income, age, born in an urban area, and year dummies (full control estimates in Appendix Table A.8).

### Heterogeneity, mechanisms, and ancillary results
- Heterogeneous effects across parental wealth ranges (Adjusted IV, Table 2):
  - Sample ranges and N:
    - 100<PW<500 (N = 205,030)
    - 500<PW<1200 (N = 50,586)
    - 1200<PW (N = 15,470)
  - Effects on wage level (preserved entries):
    - 5.137 (8.097)
    - 0.0380*** (0.0116)
    - 0.00790*** (0.00284)
  - Effects on wage percentile:
    - 0.294 (0.468)
    - 0.00226*** (0.000712)
    - 0.000519*** (0.000243)
  - Effects on wage mobility:
    - 0.867 (1.368)
    - 0.00528*** (0.00147)
    - 0.000321 (0.000374)
  - Effects on total income:
    - 15.540 (24.37)
    - 0.122*** (0.0359)
    - 0.131*** (0.0391)
  - Interpretation: effects insignificant in lower range (NOK 100,000 to 500,000); significant and largest in middle range (NOK 500,000 to 1.2 million); smaller but still significant at top range.
- Human capital channel:
  - Controlling for higher education and field of study produces slightly smaller but broadly similar estimates (Appendix Table A.9).
  - Table A.9 effect of net parental wealth on wage level controlling for education:
    - OLS: 0.00196*** (0.000310)
    - IV: 0.0133*** (0.00103)
    - Adjusted IV: 0.0117*** (0.00104)
- Wealth and income dispersion:
  - Parental wealth associated with higher dispersion of children’s labor income after controlling for individual and parents’ characteristics—interpreted as wealth enabling different risk profiles and acting as a private safety net.
  - Table A.10 (Labor Earnings Dispersion):
    - Wage income dispersion OLS: 0.0000213*** (7.77∗10−7)
    - Wage income dispersion OLSIV: 1.85∗10−6 *** (1.19∗10−7)
    - Wage income dispersion Adjusted: 0.0000135*** (9.35∗10−7)
    - Total income dispersion OLS: 2.19∗10−6*** (5.01∗10−7)
    - Total income dispersion OLSIV: 0.000136*** (7.13∗10−7)
    - Total income dispersion Adjusted: 0.0000770*** (4.46∗10−6)
  - Binned scatterplots and IV estimates (Figure 4 and Appendix Table A.10) indicate higher parental wealth associated with higher wage dispersion; similar pattern for capital income dispersion (Appendix, Figure A.2).

### Norwegian wealth tax, data context, and descriptive statistics
- Wealth tax institutional details:
  - Marginal wealth tax rate varied from a three-step progressive rate in the mid-90s—reaching a rate of 1.5 percent—to a flat rate from the early 2000s—currently set at 0.85 percent.
  - Current threshold is a net wealth above NOK 1.5 million (about USD 174,000)—doubled for married couples—compared to NOK 125,000 and NOK 150,000 for singles and married couples, respectively, in the early 1990s.
  - In 1993, about 18 percent of Norwegian taxpayers were subject to the wealth tax; in 2017 the number had dropped to 10 percent.
- Data source: Statistics Norway’s databases, including the Income Statistics for Families and Persons (Register of Tax Returns), enabling parent–child linking and tracing of income and net wealth since 1993; dataset includes education levels, field of study, place of birth, and other characteristics.
- Wealth dynamics in sample:
  - In 1993, average net financial wealth was 85 percent of the average wage. By 2017, average net financial wealth had risen to 135 percent of the average wage.
  - In 2017 the top 10 percent wealthiest owned about half of all (positive) net wealth in Norway in 2017.
- Selected summary statistics (Table A.2 and A.3, amounts in NOK 1000):
  - Individuals, 2017 (All): Wage 513,155 (372,967); Capital income 31,770 (445,701); Total income 576,325 (603,270); N for PW>0: 63,533.
  - Parents, 1993 Net wealth: 466,032 (6,420,489); Median 233,756. Net wealth, 1998: 835,655 (4,461,313); Median 387,268.
  - Wealth tax payment, 1993: 3533 (83,412); Wealth tax payment 1998-rules, 1993: 29552 (98,126); Median 000.

### Robustness, first-stage and divorce effect evidence
- First-stage IV (Table A.5):
  - Effect of instrument on net parental wealth: −21031.6*** (1997.9) ; t-value −10.53 ; F-value 447.14 ; R2 0.625.
  - Effect on parental financial wealth: −8983.5**** (902.4) ; t-value −9.96 ; F-value 309.13 ; R2 0.510.
  - Sample restrictions: 0<PW ; N: 480,971.
  - Note: Standard errors in parentheses. *p<0.10, **p<0.05, ***p<0.01.
- Direct effect of parental divorce on children (Table A.6; OLS, sample 0<PW<100; N: 161,897):
  - On wage: −20.58*** (2.221)
  - On wage percentile: −2.112*** (0.243)
  - On wage mobility: −0.951*** (0.263)
  - On total income: −42.41*** (3.007)
- Total income results (Table A.7):
  - Effect of net parental wealth on total income:
    - OLS: 0.0329*** (0.000877) ; N: 480,971
    - IV: 0.0879*** (0.0116) ; N: 480,971
    - Adjusted IV: 0.0871*** (0.0126) ; N: 270,995 (for some specifications)
  - Effect of parental financial wealth on total income:
    - OLS: 0.0352*** (0.0101)
    - IV: 0.0887*** (0.0110)
    - Adjusted IV: 0.0796*** (0.0122)

### Conclusions and policy implications
- Main conclusion: Using exogenous variation in parental net wealth from wealth tax changes, children from wealthier families tend to have higher labor income; the Norwegian wealth tax reduces that advantage and brings children’s income closer to peers from less wealthy families.
- Policy caution: The analysis does not claim the wealth tax is the only or optimal policy to influence intergenerational income inequality; normative conclusions depend on social preferences and social welfare function.
- External relevance: Results from Norway—where public provision of goods like health and education is relatively strong—suggest the “privilege effect” of parental wealth on children’s labor income might be even stronger in countries with lower provision of public goods.
- Future research directions: further study of mechanisms linking parental wealth to children’s labor income, including the role of occupational choice, risk-taking, and non-education channels.

*Source: wpiea2021085-print-pdf - References and Appendix tables and figures as presented in the source PDF.*

### References

### wpiea2021085-print-pdf - References

### I Introduction and research question
- Focus: whether a tax on parental wealth affects the labor income distribution of the next generation in Norway.
- Cohorts analyzed: born during 1978-1980.
- Timing of parental taxation exposure: parents taxed in the late 1990s (and early 2000s) when children were at advanced stages in school.
- Outcome observation window for children in adulthood: 2013-2017.
- Three primary outcomes examined:
  - the level of wage;
  - the position on the labor income distribution;
  - position on the labor income distribution relative to that of the parents (intergenerational income mobility).

### Identification strategy and empirical design
- Two sources of variation exploited:
  - changes to the wealth tax rate in the late 1990s and early 2000s; and
  - different levels of taxation of the same level of wealth depending on marital status (married joint filers had higher threshold and deduction than single filers).
- Estimation approach: IV model using tax changes (marital status differences and tax law changes) as an instrument for changes in net parental wealth.
- Concern addressed: potential violation of exclusion restriction from divorce affecting children's future income.
  - Strategy: separately estimate direct divorce effect on children’s income using a sample of taxpayers out of the scope of the wealth tax throughout the sample period, and adjust IV estimates accordingly.
  - Authors state this strategy plausibly provides an upper bound of the divorce effect.

### Main empirical findings
- Parental wealth positively associated with children’s labor income, controlling for parents’ education and incomes and individual characteristics including education.
- Point estimate of wealth effect: a net wealth of 1 million USD in Norway increases future annual wages of the children by 14,000 USD, ceteris paribus.
- Distributional impact: counterfactual analysis suggests the wealth tax made the labor income distribution less unequal—lowering the Gini coefficient by about 1 point in the sample for 2017.
- Intergenerational mobility: children from more wealthy families experience higher labor income mobility than those from less wealthy families.
- Heterogeneous effects across wealth levels:
  - Impact of the wealth tax on children’s labor income is higher at middle levels of parental wealth.
  - For the superrich, capital income diminishes the importance of employment income.
  - Low levels of parental wealth do not appear to significantly increase the chances of improving the child’s position on the labor income distribution.

### Mechanisms and ancillary results
- Education channel muted: results control for higher education and field of study of the children, and affordability/access to high quality education is facilitated by the public education system in Norway (as noted in the source).
- Evidence pointing to non-education mechanisms:
  - Parental wealth associated with higher dispersion of children’s labor income after controlling for individual and parents’ characteristics—interpreted as wealth enabling different risk profiles and acting as a private safety net.
  - Complements findings on heterogeneous returns to capital; highlights the role of heterogeneity of labor income (in addition to capital income) in driving heterogeneous total wealth returns.

### Contribution to literature
- Connects three strands:
  - empirical literature on wealth taxation (real effects on capital accumulation; behavioral responses and revenue potential);
  - intergenerational/regional income mobility literature describing patterns without linking to wealth tax;
  - literature on mechanisms of inequality of opportunity (e.g., links between parental income/wealth, access to selective schools, and long-term outcomes).
- Policy implication highlighted: the positive analysis provides empirical evidence that a wealth tax is one instrument that can lower next-generation income inequality; normative conclusions on optimal redistribution depend on social preferences and social welfare function.

*Source: wpiea2021085-print-pdf - References*

### Section V concludes.

### wpiea2021085-print-pdf - Section V concludes

### Norwegian wealth tax and data context
- Norway levies a tax on the net wealth of individuals; the marginal wealth tax rate has varied over time from a three-step progressive rate in the mid-90s—reaching a rate of 1.5 percent—to a flat rate from the early 2000s—currently set at 0.85 percent.
- The tax threshold currently is a net wealth above NOK 1.5 million (about USD 174,000)—which is doubled for married couples—compared to NOK 125,000 and NOK 150,000 for singles and married couples, respectively, in the early 1990s.
- In 1993, about 18 percent of Norwegian taxpayers were subject to the wealth tax; in 2017 the number had dropped to 10 percent.
- Data source: Statistics Norway’s databases, including the Income Statistics for Families and Persons (Register of Tax Returns), enabling parent–child linking and tracing of income and net wealth since 1993; dataset includes education levels, field of study, place of birth, and other characteristics.
- Wealth dynamics in sample: In 1993, average net financial wealth was 85 percent of the average wage. By 2017, average net financial wealth had risen to 135 percent of the average wage. In 2017 the top 10 percent wealthiest owned about half of all (positive) net wealth in Norway in 2017.

### Empirical approach and identification
- Main identification: compare wage outcomes of children of married, unmarried, and divorced parents over time, leveraging differential tax treatment of the same wealth by marital status and changes in the wealth tax rules.
- IV strategy and notation:
  - Outcome Y_i: wages.
  - Parental wealth X_i.
  - Instrument Z_i: parental divorce (change in tax treatment due to marital status change).
  - Confounder C_i: unobservables affecting wages and related to parental wealth.
- Exclusion concern and correction:
  - If divorce directly affects children's earnings (δ≠0), IV will be biased.
  - Estimate δ from a sample with parental wealth below the tax threshold (divorce does not affect their tax payments) using equation Y_j = α + β X_j + δ Z_j + e_j to identify δ.
  - Construct adjusted outcome Y_pred_i = Y_i − δ̂ Z_i and use Y_pred_i in IV second stage to identify β (equations (6) and (7)).
  - If δ̂ is larger for the low-wealth sample, approach yields an upper bound on the divorce effect; if δ is constant across samples, adjustment yields unbiased δ.
- Specifications:
  - Cohorts: three birth cohorts 1978–1980 (ages 14–16 in 1993, 19–21 in 1998, 38–40 in 2017).
  - OLS baseline (equation (8)) controls: parental wages, education, age, marital status in 1993; individual controls include age and born outside Norway; year dummies included.
  - IV instrument: change in tax payments ∆tax payment_p,i,t = taxrule_t(netw_p,i,1993) − taxrule_1993(netw_p,i,1993), exploiting tax rule changes and marital status to create exogenous variation in parental net wealth in 1998.
  - Additional control: parental marital status in 1993, conditioning on both parents alive in 1998 so variation in tax treatment is driven by divorce.
- Instrument diagnostics:
  - First-stage F−statistics and R2 support relevance; pass Stock-Yogo cutoffs (see Appendix Table A.5 as referenced).
- Additional outcomes:
  - Child’s position on the wage distribution (percentiles).
  - Intergenerational income mobility measure: child position on wage distribution relative to parents.
  - Robustness: estimates also run using total income (see Appendix Table A.7).

### Main results (summary of Table 1 and related text)
- Overall interpretation: Using exogenous changes in parental net wealth, children from wealthier families tend to have higher labor income; a wealth tax reduces that intergenerational advantage.
- Key point estimates (monetary amounts in Table 1 are measured in NOK 1000; earlier prose provides USD conversions):
  - Effect of net parental wealth on wage level (first row, total net wealth; columns 1–3):
    - OLS: 0.00271*** (standard error (0.000356))
    - IV: 0.0167*** (0.00121)
    - Adjusted IV: 0.0140*** (0.00119)
    - Prose interpretation: OLS implies a net parental wealth of USD 1 million increases future labor income of the children by USD 27,100; IV and adjusted IV estimates are USD 16,700 and USD 14,000 respectively.
  - Effect of parental financial wealth on wage level (columns 4–6):
    - OLS: 0.00255*** (0.000395)
    - IV: 0.0179*** (0.00124)
    - Adjusted IV: 0.0101*** (0.00103)
    - Prose interpretation: corresponding USD values reported as USD 25,500 (OLS), USD 17,900 (IV), USD 10,100 (adjusted IV).
  - Effect on wage percentile (second row):
    - Net parental wealth: OLS 0.000208*** (0.0000276), IV 0.00135*** (0.000104), Adjusted IV 0.00113*** (0.000107).
    - Parental financial wealth: OLS 0.000218*** (0.0000333), IV 0.00145*** (0.000105), Adjusted IV 0.000745*** (0.000890).
  - Effect on wage mobility (third row):
    - Net parental wealth: OLS 0.000649*** (0.0000638), IV 0.00317*** (0.000180), Adjusted IV 0.00295*** (0.000191).
    - Parental financial wealth: OLS 0.000633*** (0.0000869), IV 0.00335*** (0.000190), Adjusted IV 0.00206*** (0.000158).
  - Sample and sample sizes reported in Table 1:
    - Sample restrictions: 0<PW, 0<PW, 200<PW (as shown per column headings).
    - N = 480,971 (columns 1,2,4,5); N = 270,995 (columns 3,6).
- Inequality counterfactual:
  - Estimated counterfactual wage distribution in the absence of the Norwegian wealth tax (based on adjusted IV, column 3): observed Gini coefficient close 0.24; counterfactual Gini coefficient in the sample close to 0.25.
- Controls: parental wages, education and age of father and mother, parental wealth and marital status in 1993, whether the individual earns mainly capital income, age, born in an urban area, and year dummies; Appendix reports full control estimates (Table A.8).

### Extensions and mechanisms
- Heterogeneity across parental wealth ranges (Adjusted IV results, Table 2):
  - Sample ranges:
    - 100<PW<500 (N = 205,030)
    - 500<PW<1200 (N = 50,586)
    - 1200<PW (N = 15,470)
  - Effects on wage level (Table 2 entries preserved as presented):
    - 5.137 (standard error (8.097))
    - 0.0380*** (0.0116)
    - 0.00790*** (0.00284)
  - Effects on wage percentile:
    - 0.294 (0.468)
    - 0.00226*** (0.000712)
    - 0.000519*** (0.000243)
  - Effects on wage mobility:
    - 0.867 (1.368)
    - 0.00528*** (0.00147)
    - 0.000321 (0.000374)
  - Effects on total income:
    - 15.540 (24.37)
    - 0.122*** (0.0359)
    - 0.131*** (0.0391)
  - Interpretation: effects insignificant in the lower range (NOK 100,000 to 500,000); significant and largest in the middle range (NOK 500,000 to 1.2 million); smaller but still significant at the top range.
- Mechanisms:
  - Human capital channel examined: controlling for higher education and field of study produces slightly smaller but broadly similar estimates (see Appendix Table A.9).
  - Novel mechanism documented: positive correlation between net parental wealth and dispersion in returns to labor (measure: coefficient of variation of labor income across bins of parental wealth), suggesting parental wealth affects risk-taking and occupational choice.
  - IV estimates and binned scatterplots (Figure 4 and Appendix Table A.10) indicate higher parental wealth associated with higher wage dispersion; similar pattern found for capital income dispersion (Appendix, Figure A.2).
  - Interpretation linked to literature on concentration of wealth across generations (e.g., risk-taking, bequests): parental wealth as a private safety net may allow different occupational/risk choices (Pfeffer and Rodems (2021); Fagereng et al., 2021 referenced in text).

### Conclusions and policy implication highlights
- Main conclusion: Using exogenous variation in parental net wealth from wealth tax changes, children from wealthier families tend to have higher labor income; the Norwegian wealth tax reduces that advantage and brings children’s income closer to peers from less wealthy families.
- Policy caution: The analysis does not claim the wealth tax is the only or optimal policy to influence intergenerational income inequality, but provides evidence that a wealth tax can raise intergenerational income mobility.
- External relevance: Results from Norway—where public provision of goods like health and education is relatively strong—raise the question whether the observed “privilege effect” of parental wealth on children’s labor income might be even stronger in countries with lower provision of public goods.
- Future research directions: further study of mechanisms linking parental wealth to children’s labor income, including the role of occupational choice, risk-taking, and non-education channels.

*Section V concludes. Source: wpiea2021085-print-pdf - Section V concludes.*

### References

### wpiea2021085-print-pdf - References

### References
- Benhabib, J., & Bisin, A. (2018). Skewed wealth distributions: Theory and empirics. Journal of Economic Literature, 56(4), 1261–1291.
- Bjørneby, M., Markussen, S., & Røed, K. (2020). Does the wealth tax kill jobs? (Tech. rep.). IZA DP No. 13766.
- Boserup, S. H., Kopczuk, W., & Kreiner, C. T. (2018). Born with a silver spoon? danish evidence on wealth inequality in childhood. The Economic Journal, 128(612), F514–F544.
- Chetty, R., Friedman, J. N., Hendren, N., Jones, M. R., & Porter, S. R. (2018). The opportunity atlas: Mapping the childhood roots of social mobility (tech. rep.). National Bureau of Economic Research.
- Chetty, R., Friedman, J. N., Saez, E., Turner, N., & Yagan, D. (2020). Income segregation and intergenerational mobility across colleges in the united states. The Quarterly Journal of Economics, 135(3), 1567–1633.
- Chetty, R., Hendren, N., Kline, P., & Saez, E. (2014). Where is the land of opportunity? the geography of intergenerational mobility in the united states. The Quarterly Journal of Economics, 135(3), 1553–1623.
- Chetty, R., Hendren, N., Kline, P., Saez, E., & Turner, N. (2014). Is the united states still a land of opportunity? recent trends in intergenerational mobility. American Economic Review, 104(5), 141–47.
- Corak, M. (2013). Income inequality, equality of opportunity, and intergenerational mobility. Journal of Economic Perspectives, 27(3), 79–102.
- Fagereng, A., Mogstad, M., & Ronning, M. (2021). Why do wealthy parents have wealthy children? Journal of Political Economy, (Forthcoming).
- Jakobsen, K., Jakobsen, K., Kleven, H., & Zucman, G. (2020). Wealth taxation and wealth accumulation: Theory and evidence from denmark. The Quarterly Journal of Economics, 135(1), 329–388.
- Lee, C.-I., & Solon, G. (2009). Trends in intergenerational income mobility. Review of Economics and Statistics, 91(4), 766–772.
- Pfeffer, F., & Rodems, R. (2021). Avoiding material hardship: The buffer function of wealth (tech. rep.). University of Michigan.
- Piketty, T., & Zucman, G. (2014). Capital is back: Wealth-income ratios in rich countries 1700–2010. The Quarterly Journal of Economics, 129(3), 1255–1310.
- Ring, M. A. (2020). Wealth taxation and household saving: Evidence from assessment discontinuities in norway. Working paper.
- Saez, E., & Zucman, G. (2019). Progressive wealth taxation. Brookings Papers on Economic Activity.
- Scheuer, F., & Slemrod, J. (2020). Taxation and the superrich. Annual Review of Economics, 12(1), 189–211.
- Scheuer, F., & Slemrod, J. (2021). Taxing our wealth. Journal of Economic Perspectives, 35(1), 207–230.
- Seim, D. (2017). Behavioral responses to wealth taxes: Evidence from sweden. American Economic Journal: Economic Policy, 9(4), 395–421.
- Smith, M., Zidar, O., & Zwick, E. (2020). Top wealth in america: New estimates and implications for taxing the rich (tech. rep.). NBER.
- Thoresen, T. O. (2009). Income mobility of owners of small businesses when boundaries between occupations are vague. CESifo Working Paper Series No. 2633.

*References end.*

### Appendix: Further Results — key tables and statistics

- Table A.1: Thresholds and Deductions in the Norwegian Wealth Tax 1993-2017
  - Columns: Singles Threshold 1 / Threshold 2 / Threshold 3 (NOK) and Married Threshold 1 / Threshold 2 / Threshold 3 (NOK)
  - Selected year entries (NOK preserved exactly as in source):
    - 1993: Singles 120,000 / 235,000. / 150,000 ; Married 260,000.
    - 1994–1997: Singles 120,000 / 235,000 / 530,000 ; Married 150,000 / 260,000 / 570,000
    - 1998–2004: Singles 120,000 / 540,000. / 150,000 ; Married 580,000.
    - 2005: Singles 151,000 / 540,000. / 181,000 ; Married 580,000.
    - 2006: Singles 200,000 / 540,000. / 400,000 ; Married 1,080,000.
    - 2007: Singles 220,000 / 540,000. / 440,000 ; Married 1,080,000.
    - 2008: Singles 350,000 / 540,000. / 700,000 ; Married 1,080,000.
    - 2009: Singles 470,000.. / 940,000..
    - 2010: Singles 700,000.. / 1,400,000..
    - 2011: Singles 700,000.. / 1,400,000..
    - 2012: Singles 750,000.. / 1,500,000..
    - 2013: Singles 870,000.. / 1,740,00..
    - 2014: Singles 1,000,000.. / 2,000,000..
    - 2015: Singles 1,200,000.. / 2,400,00..
    - 2016: Singles 1,400,000.. / 2,800,000..
    - 2017: Singles 1,480,000.. / 2,960,000..
  - Note: Until 2006, married couples share one basic allowance and a joint threshold. From 2006, married couples share twice the threshold of singles on their total wealth. The threshold for singles and married is therefore the same independently of the distribution of couple wealth after 2006.

- Table A.2: Summary Statistics, Individuals, 2017 (All monetary amounts in NOK 1000; standard deviation in parentheses)
  - Strategy columns: All / Married 1993-98 / Unmarried and divorced 1993-98
  - Wage: 513,155 (372,967) / 526,999 (372,580) / 454,621 (368,756)
  - Capital income: 31,770 (445,701) / 31,591 (363,359) / 32,677 (694,284)
  - Total income: 576,325 (603,270) / 591,325 (540,422) / 512,873 (815,183)
  - Number of siblings: 1.94 (1.16) / 1.90 (1.11) / 2.10 (1.34)
  - Born in urban area: 0.145 (0.352) / 0.151 (0.358) / 0.119 (0.323)
  - Sample restrictions: PW>0 ; N: 63,533 (All) / 51,318 (Married) / 12,127 (Unmarried)

- Table A.3: Summary Statistics, Parents (Main Variables) (All monetary amounts in NOK 1000; standard deviation in parentheses)
  - Married 1993: 0.856 (0.351)
  - Divorce 1993-1998: 0.0552 (0.228)
  - Net wealth, 1993: 466,032 (6,420,489) ; Median 233,756
    - Married 1993-98: 462,658 (1,096,826) ; Median 252,307
    - Unmarried/divorced 1993-98: 479,031 (14,450,127) ; Median 153,497
  - Net wealth, 1998: 835,655 (4,461,313) ; Median 387,268
  - Financial wealth, 1993: 320,826 (5,956,855) ; Median 122,150
  - Financial wealth, 1998: 575,301 (6,703,878) ; Median 138,048
  - Wealth tax payment, 1993: 3533 (83,412) ; Median 000
  - Wealth tax payment 1998-rules, 1993: 29552 (98,126) ; Median 000
  - Sample restrictions: PW>0 ; N: 63,533 / 51,318 / 12,127

- Table A.4: Summary Statistics, Parents (Further Variables) (All monetary amounts in NOK 1000; standard deviation in parentheses)
  - Mother’s wage: 105,846 (83,144) ; Married 107,698 (81,344) ; Unmarried 98,093 (89,831)
  - Father’s wage: 199,051 (158,989) ; Married 213,627 (153,314) ; Unmarried 136,577 (164,258)
  - Mother’s capital income: 730 (62,510)
  - Father’s capital income: 30,340 (291,760)
  - Mother’s total income: 127,329 (117,451)
  - Father’s total income: 277,991 (372,076)
  - Mother higher education: 0.228 (0.420)
  - Father higher education: 0.173 (0.378)
  - Sample restrictions: PW>0 ; N: 63,533 / 51,318 / 12,127

- Table A.5: First-Stage IV Estimation Results (All monetary amounts in NOK 1000; standard errors in parentheses)
  - Effect of Instrument
    - On net parental wealth: -21031.6*** (1997.9) ; t-value -10.53 ; F-value 447.14 ; R2 0.625
    - On parental financial wealth: -8983.5**** (902.4) ; t-value -9.96 ; F-value 309.13 ; R2 0.510
  - Sample restrictions: 0<PW ; N: 480,971
  - Note: Standard errors in parentheses. *p<0.10, **p<0.05, ***p<0.01. PW is net parental wealth divided by the number of siblings in 1993. Controls listed in source; mobility outcomes measured in percentiles from father’s wage income in 1993 to children’s wage income in 2010-2017.

- Table A.6: Direct Effect of Divorce on Wages of Children (OLS; All monetary amounts in NOK 1000; standard errors in parentheses)
  - Effect of Parental divorce:
    - On wage: -20.58*** (2.221)
    - On wage percentile: -2.112*** (0.243)
    - On wage mobility: -0.951*** (0.263)
    - On total income: -42.41*** (3.007)
  - Sample restrictions: 0<PW<100 ; N: 161,897

- Table A.7: Total Income (Effects reported with standard errors in parentheses)
  - Effect of Net parental wealth on total income:
    - OLS: 0.0329*** (0.000877) ; N: 480,971
    - IV: 0.0879*** (0.0116) ; N: 480,971
    - Adjusted IV: 0.0871*** (0.0126) ; N: 270,995 (for some specifications)
  - Effect of Parental financial wealth on total income:
    - OLS: 0.0352*** (0.0101)
    - IV: 0.0887*** (0.0110)
    - Adjusted IV: 0.0796*** (0.0122)
  - Sample restrictions vary by column (0<PW, 0<PW200<PW, etc.)

- Table A.8: Main Results with Effect of Controls (All monetary amounts in NOK 1000; standard errors in parentheses)
  - Effects on Wage (first effect is change in net parental wealth 1993-1998 instrumented by wealth tax change; net parental wealth 1993 included)
    - Change in net parental wealth 1993-1998:
      - OLS: 0.00271*** (0.000356)
      - IV: 0.0167*** (0.00121)
      - Adjusted IV: 0.0140*** (0.00119)
    - Net parental wealth 1993:
      - OLS: 0.00133*** (0.000182)
      - IV: 0.00773*** (0.000560)
      - Adjusted IV: 0.00650*** (0.000551)
    - Parents married in 1993: 35.35*** (1.194) ; IV: 33.85*** (1.209) ; Adjusted IV: 34.74*** (1.250)
    - Earning mainly capital income 2013-2017: -442.9*** (0.930) ; IV: -449.4*** (1.153) ; Adjusted IV: -459.9*** (1.698)
    - Father’s wage 1993: 0.176*** (0.00342) ; IV: 0.142*** (0.00471) ; Adjusted IV: 0.134*** (0.00593)
    - Mother’s wage 1993: 0.179*** (0.00548) ; IV: 0.184*** (0.00567) ; Adjusted IV: 0.181*** (0.00763)
    - Born in an urban area: 15.81*** (1.252) ; IV: 16.76*** (1.258) ; Adjusted IV: 15.75*** (1.841)
    - Father has higher education: 23.37*** (1.316) ; IV: 25.06*** (1.341) ; Adjusted IV: 25.78*** (1.863)
    - Mother has higher education: 23.72*** (1.085) ; IV: 21.16*** (1.127) ; Adjusted IV: 17.64*** (1.465)
  - Sample restrictions: PW>0 ; N: 481,319 (OLS/IV) ; Adjusted IV N: 292,673

- Table A.9: Results Controlling for Educational Level and Field
  - Effect of Net parental wealth on wage level:
    - OLS: 0.00196*** (0.000310)
    - IV: 0.0133*** (0.00103)
    - Adjusted IV: 0.0117*** (0.00104)
  - Effect on wage percentile:
    - OLS: 0.000209*** (0.0000272)
    - IV: 0.00130*** (0.000101)
    - Adjusted IV: 0.00108*** (0.000104)
  - Effect on wage mobility:
    - OLS: 0.000652*** (0.0000637)
    - IV: 0.00313*** (0.000177)
    - Adjusted IV: 0.00291*** (0.000188)
  - Effect of Parental financial wealth reported similarly (exact coefficients and standard errors preserved in source).
  - Sample restrictions and N: 480,971 / 270,995 as reported.

- Table A.10: Labor Earnings Dispersion
  - Strategy: OLS (no controls) / OLSIV
  - Effect of Net parental wealth:
    - Wage income dispersion: 0.0000213*** (7.77∗10−7) ; OLSIV: 1.85∗10−6 *** (1.19∗10−7) ; Adjusted: 0.0000135*** (9.35∗10−7)
    - Total income dispersion: 2.19∗10−6*** (5.01∗10−7) ; OLSIV: 0.000136*** (7.13∗10−7) ; Adjusted: 0.0000770*** (4.46∗10−6)
  - N: 481,319
  - Note: Effect of parental wealth on dispersion calculated by constructing 100 groups of parental wealth and calculating dispersion within each group; parental wealth groups constructed beginning with a bin of 100,000 and increasing by a factor of bin^(1.1) to measure dispersion at the top.

- Figures (descriptive notes)
  - Figure A.1: Income Inequality (Considering Heterogeneous Effects of Parental Wealth on Income) — panels for (a) Wage Income and (b) Total Income.
  - Figure A.2: Total Income Dispersion and Parental Wealth Levels
    - Panel (a) Total Income Dispersion
    - Panel (b) Total Income Dispersion, Controlling for Education
    - Note: For each bin of the logarithm of net parental wealth in 1993, the figure shows the estimated relationship between net parental wealth and capital income dispersion, controlling for parents and individual’s characteristics including education. The measure of total income dispersion is the coefficient of variation defined as the ratio of standard deviation to the mean (averaged within each bin of wealth).

*Content unit: References and Appendix tables and figures as presented in the source PDF.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021085-print-pdf.pdf_