Chapter 1. Introduction

This chapter sets the scene for a detailed analysis of the taxation of savings in OECD and partner economies. The chapter examines recent trends in inequality and population ageing. It then summarises the economic literature on the impact of taxation on household savings. Finally, it provides an outline of the structure of the rest of the report.

The taxation of household savings has long been an important issue for tax policymakers. This is unsurprising because how a country taxes household savings can have significant effects on both the efficiency and fairness of the tax system. This report examines in detail the taxation of household savings in OECD and five key partner countries and discusses the resulting implications for savings tax policy.

Such a review is very timely in light of a range of trends and recent developments that have significant implications for tax policy. In particular, inequality continues to increase in many countries leading to strong calls for greater taxation of savings and wealth; while recent changes in the international tax environment regarding the exchange of financial account information between tax administrations can be expected to significantly alter the ability of countries to tax capital income. Meanwhile, continued population ageing is placing significant pressure on public pension systems, with many countries therefore looking to further encourage private pension savings.

This introductory chapter provides a range of background information to aid the analysis to follow. First, it examines recent trends in inequality and population ageing. Second, it summarises the economic literature on the impact of taxation on savings. Finally, the chapter also provides an outline of the structure of the report to follow.

1.1. Background

Inequality is increasing

Within-country income inequality has increased over recent years in many countries due to a wide range of factors including technological change and globalisation, as well as tax, benefit and other policy settings.1 These increases have been brought into particular focus as a result of the 2008 global financial and economic crisis, and have led to strong calls for greater taxation of savings and wealth in many countries (see, e.g., Piketty, 2014).

The Gini coefficient of disposable income inequality stood at 0.29 on average across OECD countries in the mid-1980s. By 2013, it had increased by about 10% or three points to 0.32 (OECD, 2015a). The latest available data shows that inequality has risen since the mid-1980s in 19 of the 22 OECD countries for which long-time series data are available (Figure 1.1).

Wealth is much more unequally distributed than income. In the 18 OECD countries for which comparable data is available, the bottom 40% of the wealth distribution own only 3% of total household wealth (Figure 1.2). In comparison, the bottom 40% of the income distribution receives 20% of total household income. The top 10% of the wealth distribution hold half of total household wealth and the wealthiest 1% own almost a fifth. The wealth share of the top 1% of the wealth distribution is almost as large as the income share of the top decile in the income distribution (OECD, 2015a).2

There is also some evidence that wealth inequality has been worsening over time. Using data from eight OECD countries since the 1970s, Piketty (2014) concluded that private wealth has tended to become more unequally distributed in recent decades, reversing a long-term decline throughout much of the 20th century.

Some caution is necessary when considering measures of income and wealth inequality – with potential for both under- and over-estimation of levels and trends in income inequality. For example, some studies have suggested that commonly used data sources – including those cited above – underestimate the levels of income inequality by not accounting for tax evasion and avoidance, and by not valuing wealth held inside businesses (Piketty and Saez, 2006). Some studies suggest that tax evasion is higher for high income and high net wealth households than for lower income and net wealth households, meaning that a larger percentage of total income and wealth go unreported at the top of the income and net wealth distribution than at the bottom (Zucman, 2014). This suggests that when concealed wealth and income is taken into account, income and wealth inequality is even greater than the many existing estimates suggest. In contrast, several studies of income inequality in the United States find lower levels and smaller increases in inequality using broader measures of income and correcting for factors such as declining marriage rates and increasing transfer payments (Auten and Splinter, 2017; Bricker et al., 2016; Burkhauser, et al., 2012).3

Additionally, trends across time – such as those illustrated in Figure 1.1 – may be affected by tax avoidance and evasion behaviour. For example, part of the increases in measured Gini inequality in the late 1980s may be due to the base broadening reforms that occurred in many countries in that period, resulting in previously undeclared income of higher-income taxpayers being brought into the tax base (and reflected to some degree in survey responses).4 The choice of start and end points in such a comparison across time may also influence trends.5 Furthermore, the causes of changes in inequality over time are likely to be important in assessing policy concern (Ball and Creedy, 2016). For example, changes in population structure, such as the increasing size of older cohorts, may increase Gini inequality. However, if the (now larger) older cohort is no worse off than previous (smaller) older cohorts, policymakers may be less concerned by the measured increase in Gini inequality.

Populations are ageing

The current discussion regarding the taxation of household savings occurs against the backdrop of the ageing of our societies. A key reason why individuals save is to provide for their retirement. Thus, the tax treatment of retirement pensions in particular needs to be considered in the context of the financing of both public and private pensions across countries.

In many OECD countries, demographic changes have led to a smaller number of workers supporting a larger number of retirees through pay-as-you-go schemes. Figure 1.3 shows the old-age dependency ratio (the number of individuals aged 65 and over per 100 people of working age) across OECD countries. In 2015, the average old-age dependency ratio across OECD countries was 27.6. Figure 1.4 shows that this ratio has been steadily increasing since 1950. Projections beyond 2015 suggest even greater increases, with more than 50 retirees per 100 people of working age on average by 2050.

OECD research shows that securing adequate pension arrangements requires a diversity of pension retirement financing. This includes public pension arrangements as well as private financing. This is particularly true in the context of high public deficits across the OECD. Diversification of public and private retirement provision allows for greater security of retirement provision, because different kinds of schemes are more or less exposed to different kinds of risks. For example, population ageing may create fiscal sustainability pressures for pay-as-you-go defined benefit arrangements, solvency problems for funded defined benefit arrangements, and adequacy problems for funded defined contribution pension arrangements (OECD, 2016).

These demographic trends have led to an increasing focus on private pension provision, and on defined contribution schemes. These factors should be considered when examining the merits of tax incentives for private pensions.

1.2. The effects of taxation on savings

A key concern for tax policymakers is the potentially distortionary impact of taxation on savings. Taxation can potentially affect savings in two ways. It may affect the allocation of savings across different assets (i.e. “portfolio composition”); and the allocation of savings over the lifecycle (and hence on the total level of savings of an individual and in an economy). This section provides a brief summary of the empirical evidence examining the effect of taxation on savings behaviour.

The effect of taxation on portfolio composition

There is a long history of empirical analysis investigating the impact of taxation on household savings portfolio composition. While these studies vary in the strength of their findings, a clear conclusion can be reached that taxpayers do respond to incentives in the tax system and alter portfolio composition towards more tax-favoured assets.

The first econometric analysis of the impact of taxation on portfolio composition was by Feldstein (1976). He found the personal income tax to have a “very powerful effect on individuals’ demands for portfolio assets, after adjusting for the effects of net wealth, age, sex, and the ratio of human to non-human capital.” However, Feldstein disregards the methodological implications of incomplete portfolios (i.e. the fact that many households will hold only a subset of the group of assets considered in a study) – a shortcoming that may lead to sample selection bias. Cross-sectional studies from the early eighties through the nineties utilise the Heckman (1979) two-step approach to overcome the methodological limitations in Feldstein’s analysis.

The Heckman two-step approach suggests that households make two decisions when making savings decisions that affect the composition of their portfolios. First, households decide whether to hold a particular asset or not (the discrete asset allocation choice). Conditional on deciding to hold an asset, the household then decides how much of that asset to hold (the continuous asset allocation choice).

Cross-sectional studies following the Heckman two-step approach include Dicks-Mireaux and King (1983) for Canada; Hubbard (1985), King and Leape (1998) and Poterba and Samwick (2002) for the United States; and Agell and Edin (1990) for Sweden. These papers consistently find that taxes have a significant effect on the decision to hold a particular asset, although evidence is more mixed on the effect of taxes on the decision as to how much of an asset to hold.

More recent studies (e.g. Samwick, 2000; Alan and Leth-Petersen, 2006; Alan et al., 2010; and Zoutman, 2014) have tended to be ‘natural experiment’ type studies, utilising time series data and the variation in tax rates following a tax reform to estimate the impact of taxes on portfolio composition. Findings from these more recent studies generally indicate that taxes have a significant effect on portfolio composition, though they find smaller effects than those suggested in cross-sectional studies.

Most studies have focused primarily on the impact of taxation on the portfolio composition of high-income and wealthy households. An exception is the field experiment of Duflo et al. (2006) where low- and middle-income households are considered. While this study did find evidence that taxes impact the asset allocation decisions of lower income households, the study concluded that factors such as information costs and framing also play an important role.

The effect of taxation on the level of savings6

Empirical evidence is vastly more mixed regarding whether taxation affects the level of savings in an economy. In this regard, most empirical analysis has focused on tax incentives for retirement savings.

OECD (2018) undertakes a detailed review of the literature since the mid-1990s, and finds two disparate groupings of papers.7 A first set of papers generally finds that tax incentives for retirement savings produce at least moderate increases in national savings. For example, Poterba et al. (1996) find that the tax incentives associated with IRA and 401(k) plans in the United States have led to increased national savings. Other papers with similar findings include: Hubbard and Skinner (1996); Benjamin (2003), Ayuso et al. (2007), Guariglia and Markose (2000), Rossi (2009), Gelber (2011) and Beshears et al. (2014).

In contrast, a similarly large group of studies finds that tax incentives for retirement savings mostly result in reallocation rather than new savings. Engen et al. (1996), for example, who also consider IRAs in the United States, find instead that most of the reported increase in financial assets in IRAs can be attributed to stock market booms, higher real interest rates, and shifts in non-financial assets, debt, pensions and social security wealth. Papers with similar findings include: Attanasio and DeLeire (2002), Pence (2002), Attanasio et al. (2004), Antón et al. (2014), Chetty et al. (2014) and Paiella and Tiseno (2014).

Interestingly, there is more consistent evidence that tax incentives tend to increase the level of savings of low-income individuals, while high-income individuals tend to reallocate their savings (Engelhardt, 2000; Engen and Gale, 2000; Chernozhukov and Hansen, 2004; Engelhardt and Kumar, 2011).

1.3. Outline of the report

The rest of this report is structured as follows. Chapter 2 investigates how countries actually tax savings, finding wide variation in approaches both across and within countries, and across different asset types. Broad trends in design approaches are examined and a summary of the key design features adopted by countries is provided.

Chapter 3 quantifies the incentives created by these tax systems by presenting marginal effective tax rates (METRs) on a wide range of assets / savings vehicles. METR modelling enables the impact of a wide range of taxes and tax design features to be incorporated into the one indicator. The results for the 40 countries examined highlight significant variation in METRs across savings vehicles, with tax systems creating significant incentives to alter portfolio allocation away from that which would be optimal in the absence of taxation, with some assets tending to be particularly tax-favoured as compared to others.

Chapter 4 analyses patterns of asset holdings in selected OECD countries across both income and wealth distributions. The analysis draws on income and asset-holdings microdata for 18 countries from the Eurosystem Household Finance and Consumption Survey (HFCS). Patterns of asset holdings are found to vary substantially across both income and wealth distributions, with significant implications regarding the distributional effects of the taxation of household savings.

Chapter 5 discusses the taxation of household savings from an international perspective, and considers the impact of exchange of information on the taxation of capital income internationally. The recent move towards the automatic exchange of financial account information between tax administrations suggests that it will be harder in years to come for taxpayers to engage in tax evasion, and thus will make it easier to levy taxes on capital income without income and assets necessarily shifting offshore in response.

Finally, Chapter 6 brings together the key insights from the preceding chapters and discusses their implications for savings tax policy. It concludes that, while countries do not necessarily need to tax savings more, there is significant scope to improve the way countries tax savings. Most significantly, there are opportunities for countries to increase coherency and consistency in taxation across assets and thereby improve both the efficiency and fairness of their tax systems. There may also be opportunities for many countries to increase progressivity in their taxation of savings. At the same time, there remains a case for well-designed preferential tax treatment of private pensions in order to encourage retirement savings.

References

Agell, J. and P. Edin (1990), “Marginal taxes and the asset portfolios of Swedish households”, Scandinavian Journal of Economics, Vol. 92/1, Wiley, pp. 47-64.

Alan, S., K. Atalay, T. Crossley and S-H. Jeon (2010), “New evidence on taxes and portfolio choice”, Journal of Public Economics, Vol. 94/11-12, Elsevier, pp. 813-823.

Alan, S. and S. Leth-Petersen (2006), “Tax Incentives and Household Portfolios: A Panel Data Analysis”, SEDAP Research Paper, No. 163, McMaster University.

Antón, J., R. Muñoz de Bustillo and E. Fernández-Macías (2014), “Supplementary private pensions and saving: evidence from Spain”, Journal of Pension Economics and Finance, Vol. 13/4, Cambridge University Press, pp. 367-388.

Attanasio, O., J. Banks and M. Wakefield (2004), “Effectiveness of tax incentives to boost (retirement) saving: Theoretical motivation and empirical evidence”, IFS Working Paper, W04/33.

Attanasio, O. and T. DeLeire (2002), “The effect of individual retirement accounts on household consumption and national saving”, The Economic Journal, Vol. 112/481, Wiley, pp. 504-538.

Auten, G. and D. Splinter (2017), “Income inequality in the United States: Using tax data to measure long-term trends”, US Treasury Department and Joint Committee on Taxation (unpublished paper)

Ayuso, J., J. Jimeno and E. Villanueva (2007), “The effects of the introduction of tax incentives on retirement savings”, Bank of Spain Working Papers, No. 0724, Banco de España, Madrid,.

Ball, C. and J. Creedy (2016), “Inequality in New Zealand 1983/84 to 2012/13”, New Zealand Economic Papers, Vol. 50/3, pp. 323-342.

Benjamin, D. (2003), “Does 401(k) eligibility increase saving? Evidence from propensity score subclassification”, Journal of Public Economics, Vol. 87/5-6, Elsevier, pp. 1259-1290.

Beshears, J., J. Choi, D. Laibson and B. Madrian (2014), “Does front-loading taxation increase savings? Evidence from Roth 401(k) introductions”, NBER Working Paper Series, No. 20738, The National Bureau of Economic Research.

Bricker, J., A. Henriques, J. Krimmel and J. Sabelhaus (2016), “Measuring Income and Wealth at the Top Using Administrative and Survey Data”, Brookings Papers on Economic Activity, Brookings, pp. 261-312.

Burkhauser, R., S. Feng, S. Jenkins and J. Larrimore (2012), “Recent Trends in Top Income Shares in the United States: Reconciling Estimates from March CPS and IRS Tax Return Data”, The Review of Economics and Statistics, Vol. 44/2, pp. 371-388.

Chernozhukov, V. and C. Hansen (2004), “The effect of 401(k) participation on the wealth distribution: An instrumental quantile regression analysis”, The Review of Economics and Statistics, 86/3, pp. 735-751.

Chetty, R., J. Friedman, S. Leth-Petersen, T. Nielsen and T. Olsen (2014), “Active vs. passive decisions and crowdout in retirement savings accounts: Evidence from Denmark”, The Quarterly Journal of Economics, Vol. 129/3, pp. 1141-1219.

Congressional Budget Office (2016), The Distribution of Household Income and Federal Taxes, 2013, (supplemental tables) Congressional Budget Office, www.cbo.gov/publication/51361 .

Dicks-Mireaux, L. and M. King (1983), “Portfolio Composition and Pension Wealth: An Econometric Study”, in Z. Bodie and J. Shoven (eds.), Financial Aspects of the United States Pension System, University of Chicago Press.

Duflo, E., W. Gale, J. Liebman, P. Orszag and E. Saez (2006), “Saving Incentives for Low- and Middle-Income Families: Evidence from a Field Experiment with H&R Block”, The Quarterly Journal of Economics, Vol. 121/4, pp. 1311-1346.

Engelhardt, G. (2000), “Have 401(k)s raised household saving? Evidence from the Health and Retirement Study”, Syracuse Centre for Policy Research Working Paper.

Engelhardt, G. and A. Kumar (2011), “Pensions and household wealth accumulation”, The Journal of Human Resources, Vol. 46/1.

Engen, E. and W. Gale (2000), “The effects of 401(k) plans on household wealth: Differences across earnings groups”, NBER Working Paper Series, No. 8032.

Feldstein, M. (1976), “Personal taxation and portfolio composition: an econometric analysis”, Econometrica, Vol. 44/4, JSTOR, pp. 631–650.

Gelber, A.M. (2011), “How do 401(k)s affect saving? Evidence from changes in 401(k) eligibility”, American Economic Journal: Economic Policy, American Economic Association, Vol. 3/4, pp. 103-122.

Guariglia, A. and S. Markose (2000), “Voluntary contributions to personal pension plans: Evidence from the British Household Panel Survey”, Fiscal Studies, Vol. 21/4, pp. 469-488.

Heckman, J. (1979), “Sample Selection Bias as a Specification Error”, Econometrica, Vol. 47/1, JSTOR, pp. 153-161.

Hubbard, R. (1985), “Personal taxation, pension wealth, and portfolio composition”, Review of Economics and Statistics, Vol. 67/1, JSTOR, pp. 53-60.

Hubbard, R. and J. Skinner (1996), “Assessing the effectiveness of saving incentives”, Journal of Economic Perspectives, American Economic Association, Vol. 10/4, pp. 73-90.

King, M. and J. Leape (1998), “Wealth and portfolio composition: theory and evidence”, Journal of Public Economics, Vol. 69/2, pp. 155-193.

OECD (2018), OECD Pensions Outlook 2018, OECD Publishing, Paris (forthcoming).

OECD (2016), OECD Pensions Outlook, 2016. OECD Publishing, Paris, https://doi.org/10.1787/pens_outlook-2016-en .

OECD (2015a), In It Together: Why Less Inequality Benefits All, OECD Publishing, Paris, https://doi.org/10.1787/9789264235120-en .

OECD (2015b), Pensions at a Glance 2015: OECD and G20 indicators, OECD Publishing, Paris, https://doi.org/10.1787/pension_glance-2015-en .

OECD (1994), “Taxation and Household Saving”, OECD Publishing, Paris.

Paiella, M. and A. Tiseno (2014), “Evaluating the impact on saving of tax-favored retirement plans”, Journal of Pension Economics and Finance, Vol. 13/1, pp. 62-87.

Pence, K. (2002), “401(k)s and household saving: New evidence from the Survey of Consumer Finances”, Finance and Economics Discussion Series, Vol. 2002/6, Board of Governors of the Federal Reserve System, Washington, D.C.

Piketty, T. (2014), Capital in the 21^st century, Harvard University Press.

Piketty, T. and E. Saez (2006), “The Evolution of Top Incomes: A Historical and International Perspective”, American Economic Review, Papers and Proceedings, Vol. 96/2, pp. 200-205.

Poterba, J. and A. Samwick (2002), “Taxation and household portfolio composition: US evidence from the 1980s and 1990s”, Journal of Public Economics, Vol. 87/1, pp. 5-38.

Poterba, J., S. Venti and D. Wise (1996), “How retirement saving programs increases saving”, The Journal of Economic Perspectives, Vol. 10/4, pp. 91-112.

Rossi, M. (2009), “Examining the interaction between saving and contributions to personal pension plans: Evidence from the BHPS”, Oxford Bulletin of Economics and Statistics, Vol. 71/2, pp. 253-271.

Saez, E. and G. Zucman (2016), “Wealth inequality in the United States since 1913: Evidence from capitalized income tax data”, Quarterly Journal of Economics, Vol. 131/2, pp. 519-578.

Samwick, A. (2000), “Portfolio responses to taxation: Evidence from the end of the rainbow”, in: J. Slemrod (ed.), Does Atlas Shrug, Harvard University Press, Cambridge, MA.

Scholz, J. (1994), “Portfolio choice and tax progressivity: Evidence from the Surveys of Consumer Finances”, in: J. Slemrod (ed.), Tax Progressivity and Income Inequality, Cambridge University Press, New York.

Zoutman, F. (2014), “The Effect of Capital Taxation on Households’ Portfolio Composition and Intertemporal Choice: Evidence from the Dutch 2001 Capital Income Tax Reform”, NHH Dept. of Business and Management Science Discussion Paper, No. 2014/23, Norwegian School of Economics.

Zucman, G. (2014), “Taxing across Borders: Tracking Personal Wealth and Corporate Profits”, Journal of Economic Perspectives, Vol. 28/4, pp. 121-148.

Notes

← 1. An analysis of the various causes of changes in income inequality over time is beyond the scope of this report.

← 2. Commonly used measures of wealth generally include only private wealth. Most developed countries have public pension systems that provide wealth in the form of the capitalised value of future pensions. For example, in the United States the formula is based on the highest 35 years of qualified earnings so that each year of work adds to this form of wealth. Social Security wealth represents a major form of wealth for much of the population and reduces the perceived need to accumulate private wealth for retirement. While researchers note the importance of such wealth (e.g., Bricker et al., 2016), conventional wealth measures generally do not include it.

← 3. Saez and Zucman (2016) estimate that the share of total household wealth owned by the top 0.1% in the United States increased from 7% in the late 1970s to 22% in 2012. In contrast, Bricker et al. (2016) conclude that top wealth shares are lower and growing more slowly in the United States. For example, their preferred estimate is that the top 0.1 percent’s wealth share increased from about 11 percent in 1992 to 15 percent in 2013. Bricker et al. conclude that all of the difference in estimated growth of the top 0.1% share is due to the gross capitalisation rate used for fixed income assets in Saez and Zucman (2016). This capitalisation rate generated the result that fixed income assets (bank accounts and bonds) accounted for nearly half of the total assets of the top 0.1% in 2013 and virtually all of the increase in the top 0.1% share between 2001 and 2013.

← 4. Auten and Splinter (2017), for example, find base broadening in the 1980s to have had a significant impact on income reported in tax returns leading to overestimation of increases in the share of income held by the top 1% in the United States by studies that fail to correct for changes in the tax base.

← 5. For example, while survey data for New Zealand shows a higher disposable income Gini coefficient in 2013 than in 1984, much of the increase occurred in the late 1980s – such that the Gini is lower in 2013 than in 1993 (Ball and Creedy, 2016).

← 6. This section draws heavily on OECD (2018)

← 7. OECD (1994) finds similar mixed evidence from earlier studies.