Tax Expenditures and Evaluations 2012
In order to reduce the taxation of savings and improve incentives for Canadians to save, Budget 2008 introduced the Tax-Free Savings Account (TFSA)—a flexible, registered, general‑purpose account that allows Canadians to earn tax-free investment income. Starting in 2009, Canadian residents age 18 and older have been eligible to contribute to a TFSA.
The TFSA has already become a popular savings vehicle. By the end of 2011, approximately 8.2 million Canadians had opened a TFSA, and financial assets held in TFSAs were valued at over $62 billion.
This paper uses administrative account data for the first three years of the program to analyze the profile of individuals who have opened a TFSA, focusing primarily on participation and contribution trends. This exercise is useful in that it allows for a better understanding of the short-term impact of the program and of its beneficiaries, which provides some early indications of the longer-term benefits of the program. An empirical assessment of the long-term impact of TFSAs on savings behaviour would require much longer time series on individual savings and other economic variables following the introduction of the TFSA in 2009, and is therefore not within the scope of this paper.
Some salient points emerging from the analysis are:
- TFSAs are a popular means of savings for individuals across all income levels. In particular, individuals with incomes below $80,000 accounted for about 80% of all TFSA holders and TFSA contributions in 2011.
- Overall, over 30% of adult tax filers had a TFSA in 2011. With respect to age, TFSA participation rates are relatively stable between ages 25 and 49, and generally increase with age thereafter, with take-up among seniors being especially strong.
- Low-income seniors have also been taking advantage of the TFSA. In 2011, Guaranteed Income Supplement (GIS) recipients represented about 6% of TFSA holders, and their TFSA participation rate was 23%—3 percentage points higher than that of low-income individuals in general.
The strong initial take-up of the TFSA indicates that it is already a key component of Canadians’ personal finance planning. Growth in the use of TFSAs is expected to continue to be strong over the short- to medium-term as more investors learn of the benefits of saving in a TFSA, and to eventually stabilize as the program matures.
This paper is organized as follows. The first section describes the TFSA, including the principal rules governing the accounts. The second section provides summary statistics, highlighting how the TFSA has matured in its first three years of existence. The third section analyzes the profile of TFSA holders, focusing on participation rates and the distribution of account holders and contributions. The fourth section examines trends in unregistered savings since the introduction of the TFSA. The final section presents the conclusions of the review.
Description of the TFSA
The TFSA, which came into effect at the beginning of 2009, is a flexible, registered, general-purpose account that allows Canadians 18 years of age and older to earn tax-free investment income, including interest, dividends and capital gains. TFSAs are widely available to Canadians through banks, credit unions, life insurance companies and trusts. A social insurance number is required to register an account.
Contributions to a TFSA are not tax-deductible but investment income earned in a TFSA and withdrawals are tax-free. The tax assistance provided by the TFSA is, in many ways, a mirror image of that provided through the Registered Retirement Savings Plan (RRSP), for which contributions are tax-deductible but both the contributions and the investment earnings are taxable upon withdrawal.
When the TFSA was introduced, the annual contribution limit was set at $5,000 per individual, indexed to inflation in $500 increments. The annual contribution limit was $5,000 from 2009 to 2012. Due to indexation, the TFSA annual contribution limit increased to $5,500 starting in 2013.
An individual’s unused TFSA contribution room is carried forward and accumulates in future years. In addition, the full amount of withdrawals is added to the individual’s contribution room for the following calendar year to ensure that there is no loss in a person’s total tax-free savings room.
In recognition of the fact that couples often make their savings decisions and plan for their financial security on a joint basis, individuals can provide funds to their spouse or common-law partner for them to invest in their TFSA. TFSA assets are generally transferable to the TFSA of a spouse or common-law partner upon death.
The TFSA provides improved savings opportunities for low- and modest-income individuals because, in addition to the tax savings, neither the income earned in a TFSA nor withdrawals from it affect eligibility for federal income-tested benefits and credits, such as the Canada Child Tax Benefit, the Goods and Services Tax/Harmonized Sales Tax (GST/HST) Credit, the Age Credit, and Old Age Security and GIS benefits.
The TFSA also provides seniors with a tax-assisted savings vehicle to meet any ongoing savings needs, even after they reach the age of 71 and are required to convert their registered retirement savings into a retirement income vehicle.
A TFSA is generally permitted to hold the same investments as an RRSP. The RRSP qualified investment rules accommodate a broad range of investments including mutual and exchange-traded funds, publicly traded securities, government and corporate bonds, guaranteed investment certificates (GICs) and savings deposits.
Summary of the TFSA Data
To enable the Canada Revenue Agency (CRA) to determine contribution room and monitor compliance, TFSA issuers (i.e., the financial institutions administering the accounts) are required to file annual information returns. These returns include total contributions, total withdrawals, and the fair market value of the accounts at the end of the year. Financial institutions do not report to the CRA on the type or composition of assets held in TFSAs or on the investment income earned in the accounts, although the latter can be estimated.
Final data are available for years 2009 and 2010. For 2011, the analysis presented in this paper reflects preliminary TFSA account information records available as of June 2012.
By the end of 2011, approximately 8.2 million Canadians had opened a TFSA compared to about 6.7 million and 4.9 million at the end of 2010 and 2009, respectively (Table 1).
|Number of accounts1 (millions)||5.3||7.8||9.8|
|Number of individuals with a TFSA (millions)||4.9||6.7||8.2|
|Total annual contributions ($ millions)||19,063||25,402||30,711|
|Average contribution per TFSA holder ($)||3,926||3,769||3,727|
|Total annual withdrawals ($ millions)||1,957||4,957||8,128|
|Average withdrawal per TFSA holder ($)||403||736||986|
|Total end-of-year fair market value ($ millions)||18,243||40,707||62,006|
|Average end-of-year fair market value per TFSA holder ($)||3,757||6,039||7,525|
|Estimated investment income/loss2 ($ millions)||1,137||2,019||-1,285|
|1 An individual may hold more than one TFSA, similar to other tax-assisted savings vehicles.
2 Although TFSA holders sustained a net investment loss at the aggregate level in 2011, about three-quarters of TFSA holders had positive investment income in that year.
Total annual contributions to TFSAs have been on an upward trend, increasing from about $19 billion in 2009 to over $25 billion in 2010 to over $30 billion in 2011. Over the same three-year period, annual RRSP contributions averaged about $34 billion. Therefore, only three years after its introduction, the TFSA has approached the RRSP in terms of contribution flows even though TFSA contributions are drawn from after-tax dollars whereas RRSP contributions are made from pre-tax dollars.
Total TFSA annual contributions have increased over time since in each new year there are both additional contributions from individuals who contributed to TFSAs in prior years and contributions from new participants who can use their annual limits in addition to their carried-forward contribution room.
The average contribution per TFSA holder has varied slightly each year, ranging from $3,926 in 2009 to $3,727 in 2011. In general, it is expected that average annual contributions would be somewhat higher in the early years of the program since TFSA holders can finance their contributions from both their existing stock of unsheltered savings—which individuals would seek to gradually shift into TFSAs—and the flow of new savings for the year.
The average withdrawal has more than doubled from $403 in 2009 to $986 in 2011. The increase in withdrawals reflects the accumulation of funds in the accounts as well as the flexibility of the TFSA rules, which allow individuals to withdraw funds at any time and re-contribute the funds beginning the following year.
A strong flow of positive net contributions has led to a steady increase in assets held in TFSAs. At the end of 2011, accounts held $62 billion in assets as measured by their fair market value (FMV), which translates into an average end-of-year FMV of $7,525 per TFSA holder. By comparison, RRSPs, which have been in existence since 1957, held assets valued at about $842 billion at the end of 2011. The TFSA regime is still in its infancy such that its total asset value should continue to steadily increase as it matures.
It is estimated that, in aggregate, individuals earned about $1.1 billion and $2.0 billion of investment income in their TFSAs in 2009 and 2010, respectively. In 2011, account holders sustained a net aggregate loss of about $1.3 billion in their TFSAs. This loss, which is due to realized and unrealized capital losses, reflects the relatively poor performance of Canadian and global equity markets over the 2011 calendar year. Since many account holders principally hold risk-free assets such as savings deposits or GICs, about three-quarters of account holders had positive investment income in 2011.
Aggregate TFSA investment income is the key component for estimating the TFSA tax expenditure. A description of the methodology employed to calculate the TFSA tax expenditure is reported in the annex. The tax expenditure figures for the TFSA are reported in Table 1 in Part 1 of this publication.
TFSA Participation Rates and Profile of Account Holders
While the TFSA data can be used to compute summary statistics, they do not provide information on the profile of TFSA holders. The profile of TFSA holders and the participation rates are analyzed using income tax data.
TFSA Participation Rates
Over the first three years of the program, a broad base of Canadians took advantage of the TFSA. Overall, the number of TFSA holders as a proportion of adult tax filers (or the participation rate) has risen over time, increasing from 19% in 2009 to 26% in 2010 and to 31% in 2011.
In 2011, TFSA participation rates among the provinces and territories ranged from 15% to 36%. TFSA participation rates are highest in British Columbia, Alberta and Ontario (Chart 1).
TFSA participation rates were 33% for females and 29% for males in 2011. Given the positive relationship between age and TFSA participation (see below), one factor influencing the higher female participation rate could be differences in age composition between the genders which reflect the higher life expectancy of females. In addition, as noted above, individuals can provide funds for their spouse’s or common-law partner’s TFSA contribution. Couples seem to be taking advantage of this option—about 162,000 spouses or common-law partners (of which almost 80% were female) made TFSA contributions in 2011 that were greater than their individual incomes. Thus, transfers of funds between spouses and common-law partners likely increase the female participation rate above what it otherwise would have been.
TFSA participation rates are relatively stable between ages 25 and 49, and generally increase with age thereafter (Chart 2). Below age 50, TFSA participation rates peak in the 25-29 age group, before declining slightly over the next three age groups. The take-up of TFSAs for tax filers below age 30 is noteworthy given that younger adults generally have lower incomes and are less likely to hold financial assets than their older counterparts. On the other hand, for many young adults, the mid- to late-20s represent a good time to save for future major purchases or life events. After age 49, TFSA participation rates generally increase with age, for instance, rising from 28% for individuals 45 to 49 years of age to 41% for individuals 65 to 71 years of age in 2011. Among individuals aged 72 and over, the TFSA participation rate declines slightly, and is about one percentage point lower than for the 65-71 age group in 2011. Overall, seniors have been the largest users of TFSAs, with a take-up rate of 40% in 2011. Although many seniors are on a fixed income with a limited capacity to save on an ongoing basis, they have had more time to accumulate wealth and are generally well-placed to redirect their stock of existing savings to tax-assisted accounts such as the TFSA.
TFSA participation rates steadily rise with total income, increasing from 20% for individual tax filers with less than $20,000 of annual income to 58% for individual tax filers with more than $200,000 of annual income in 2011 (Chart 3). This is consistent with the general findings on the usage of tax-assisted savings accounts in member countries of the Organisation for Economic Co‑operation and Development (OECD), where participation rates generally increase with income. While participation rates have risen most for high-income individuals in absolute terms over the first three years of the TFSA, on a proportionate basis participation rates grew similarly for most income groups. The strongest proportional increase since 2009 was observed for those earning less than $20,000, with their TFSA participation rate nearly doubling from about 11% in 2009 to 20% in 2011.
Profile of TFSA Holders
Females accounted for about 55% of TFSA holders and total contributions in 2011. The higher proportion of female account holders is consistent with their higher TFSA participation rate and the fact that females account for about 52% of adult tax filers.
TFSA holders and the amount of contributions made by individuals are distributed similarly across all age groups (Chart 4). In particular, seniors (the total of the last two age categories) accounted for 26% of all TFSA holders and 32% of total contributions in 2011. Of this group, individuals age 72 years and older—who previously had only limited access to tax-assisted savings plans because they were ineligible to contribute to an RRSP—accounted for 15% of account holders and nearly 20% of total contributions.
The TFSA is also a popular means of savings for individuals across all income levels (Chart 5). For 2011, low- and middle-income earners (i.e., individuals with total incomes below $80,000 and who account for 88% of adult tax filers) comprised 82% of all TFSA holders and made 79% of all contributions. Low- and modest-income individuals (i.e., individuals with total incomes below $40,000) accounted for 49% of all account holders and 46% of total contributions.
The large portion of account holders and contributions among low- and modest-income individuals displayed in Chart 5 contrasts with the relatively lower participation rate of these individuals (Chart 3). This results from the fact that there are proportionally more individuals in these income categories.
As the TFSA matures and individuals transfer their existing unsheltered financial assets to a TFSA, it is expected that an increasing proportion of account holders will contribute an amount less than the maximum annual contribution limit. For instance, about one-third of account holders with less than $20,000 of total income contributed less than the $5,000 annual contribution limit in 2009. The portion of account holders in this income range contributing less than $5,000 increased to 64 per cent by 2011 (Chart 6). More generally, while the portion of TFSA holders who contributed less than $5,000 has been increasing over time for all income groups, this portion decreases as income rises. As a result, going forward, the proportion of contributions made by lower-income Canadians is expected to decline as TFSA contribution room continues to accumulate.
In analyzing the distribution of TFSA holders and contributions among income groups, it is important to recognize that some TFSA holders reporting a low income for tax purposes may reside in a higher-income household, in which case their individual incomes may not be truly representative of their financial situation or capacity to contribute to a TFSA. To assess the extent to which this is the case, the contributions of married or common-law TFSA holders with less than $20,000 of individual income were examined according to household income. For these spouses, the distribution of contributions according to household income shows that the majority reside in low- to modest-income households. About one-half of TFSA contributions made by spouses earning less than $20,000 originated from households with less than $40,000 in total income, and more than 80% originated from households with less than $80,000 in total income (i.e., households with an average income of $40,000 per spouse).
The TFSA and GIS Recipients
The TFSA improves savings incentives for low- and modest-income individuals because investment income earned in a TFSA and withdrawals from it are not included in income for the purposes of determining federal income-tested benefits and credits such as the Canada Child Tax Benefit, the GST/HST Credit, the Age Credit, and Old Age Security and GIS benefits. For low-income seniors, the exclusion of TFSA income and withdrawals for the purposes of calculating GIS benefits is particularly beneficial. Indeed, for GIS recipients, the TFSA provides a net rate of return equal to the pre-tax rate of return whereas the net rate of return on unregistered savings is reduced by the 50% GIS reduction rate.
At the end of 2011, about 440,000 GIS recipients held approximately $4.3 billion in assets in their TFSAs. In 2011, GIS recipients represented about 6% of TFSA holders, they accounted for about 7% of total TFSA assets, and their TFSA participation rate was 23%—3 percentage points higher than the 20% participation rate for the cohort of individuals earning less than $20,000 a year (which would include most GIS recipients). Going forward, the TFSA will continue to provide a favourable tax environment for incenting additional savings among low-income individuals, including prospective GIS recipients.
The TFSA and Trends in Unregistered Savings
As the TFSA matures, it is estimated that, by 2030, in combination with other registered savings accounts, it will permit over 90% of Canadians to hold all their financial assets in tax-efficient savings vehicles. In 2006 through 2008, the three years immediately preceding the introduction of the TFSA, the percentage of total adult tax filers reporting taxable interest and dividends ranged from 35% to 37% (Table 2). In 2009, the first year of the TFSA, the percentage of adult tax filers reporting this type of income fell to 33%. The number of tax filers reporting this type of income declined 3 percentage points to 30% in 2010 and fell a further percentage point to 29% in 2011. Certainly, numerous factors other than the TFSA may have influenced the number of individuals receiving taxable interest and dividends in these years, including volatile capital markets during the period, but the introduction of the TFSA in 2009 would seem to be an important change affecting the percentage of tax filers reporting this type of income on their tax returns.
|1 Excluding ineligible dividend income.
2 Preliminary 2011 T1 data.
Source: T1 data, 2006–2011.
Ensuring that the tax system provides meaningful incentives to save supports a more efficient allocation between current and future consumption. In particular, the accumulation of personal savings allows Canadians to improve their living standards and better align income and consumption when planning for important life events such as retirement. In addition, personal savings can provide individuals with a private safety net in case adverse circumstances such as job loss or illness cause an unexpected drop in income. More generally, savings contribute to economic growth by increasing the funds available for capital investment, which leads to a higher capacity to produce goods and services.
The evidence from the first three years of the TFSA program shows that Canadians have taken advantage of this new savings opportunity. The initial take-up was strong and the total annual contributions and assets have been growing at a steady pace. The TFSA is a popular means of saving for Canadians of all ages and income levels. In 2011, only three years after its introduction, close to one-third of adult tax filers had a TFSA. Seniors had the highest participation rate, with a take-up rate of 40%, while low- and middle-income earners comprised over 80% of TFSA holders and were responsible for about 80% of all contributions.
As the TFSA system matures and investment income compounds tax-free, TFSA funds will comprise a progressively larger portion of private savings. The early evidence shows that the TFSA has already become a key component of Canadians’ personal finance planning and an important building block in the formation of a fair, competitive and efficient personal income tax system for all Canadians.
Annex: Description of TFSA Tax Expenditure Methodology
TFSA data collected by the Canada Revenue Agency has allowed for the development of a methodology tailored to the tax expenditure framework.
In general, a tax expenditure represents, holding all else equal, the tax revenue that would have been collected but for a particular tax measure which deviates from the benchmark system. In the case of the TFSA, the tax expenditure represents the revenue that is not collected from investment income that is earned in the accounts. The tax expenditure figure does not necessarily represent the cost to the Government. For instance, the TFSA tax expenditure does not include the impact of increased GIS benefits (since the payment of additional non-taxable GIS benefits on account of the TFSA does not affect tax revenues) nor does it account for individuals substituting TFSA contributions in place of RRSP contributions (since tax expenditures are calculated on the premise that the underlying tax base would not be affected by the removal of a particular measure). Increased GIS benefits and RRSP substitution would increase and decrease the cost to the Government for a particular year, respectively.
In addition to estimates of the aggregate investment income earned in the accounts, estimating the TSFA tax expenditure requires knowledge of the tax rate on TFSA holders and of the distribution of the types of investment income earned in the accounts (interest, dividends and capital gains).
The marginal tax rate of account holders is calculated based on taxable income, age and province of residence. The federal average marginal tax rate (MTR) of TFSA holders was about 19% in 2009, 2010 and 2011. The average MTR is applied to interest income and is adjusted for the half-inclusion of capital gains and the federal Dividend Tax Credit.
Unfortunately, there is limited information available with regard to the distribution of investment income within TFSAs. The most relevant information is based on account data provided by financial institutions to Investor Economics, an economic consulting firm. These account data provide some indication of the breakdown of assets held by banks vis-à-vis brokerage firms as well as some information on the types of assets held at banks. While there is no breakdown by asset class for brokerages and financial advisors, it is reasonable to assume that these accounts hold mainly equities and bonds, either directly or through mutual funds or other fund structures.
Based on this information and making assumptions regarding rates of return for different asset classes, the amount of interest and dividend income can be estimated. Capital gains (or losses) can then be determined residually by subtracting interest and dividend income from the total investment income. For simplicity, and given that it would be unreasonable to assume that all capital gains (losses) are realized in the initial years of the program, it is currently assumed that one-fifth of capital gains (losses) are realized at the end of each year.
Bernheim, B. D., (2002). “Taxation and Saving,” Handbook of Public Economics, Vol. 3, A. J. Auerbach and M. Feldstein, editors. Elsevier Science Publishers B.V. (North Holland), Netherlands, pp. 1173-1249.
HM Revenue and Customs (2007). “Individual Attitudes to Saving: Effect of ISAs on People’s Saving Behaviour—Research into Attitudes and Motivations for Saving in ISAs,”working paper (www.hmrc.gov.uk/research/working-paper-isa.pdf).
Joulfaian, D. and D. Richardson (2001). “Who Takes Advantage of Tax-Deferred Saving Programs? Evidence from Federal Income Tax Data,” National Tax Journal, 54(3), pp. 669-688.
Organisation for Economic Co-operation and Development (2007). Encouraging Savings through Tax‑Preferred Accounts, OECD Tax Policy Studies, no. 20.
 While the comparison provides perspective on the relative significance of the TFSA in terms of personal savings, it is important to note that the base of potential contributors is larger for the TFSA than the RRSP since all residents aged 18 and over accumulate TFSA contribution room, whereas individuals may contribute to their RRSPs only to the extent that they earn income to generate contribution room. In addition, individuals cannot contribute to their RRSPs beyond the calendar year in which they turn 71 (or, for contributions to spousal or common-law partner RRSPs, beyond the calendar year in which their spouse or common-law partner turns 71). Also, the rules for establishing maximum annual contribution limits differ between the two savings vehicles.
 Statistics Canada, Pension Satellite Account, 2011. This figure also includes assets held in Registered Retirement Income Funds and locked-in retirement accounts.
 While investment income earned on assets held in a TFSA is not reported by financial institutions, it can be estimated by subtracting net annual contributions from the annual change in the fair market value of the assets. The investment returns include interest and dividend income, as well as net realized and unrealized capital gains and losses.
 In general, if an individual transfers property to their spouse or common-law partner for proceeds below the fair market value of the property, the Income Tax Act attribution rules treat any income earned on that property as income of the individual.
 For some of these spouses, the high level of contributions relative to income could represent a transfer of existing assets held by the spouse into a TFSA. However, the majority of the spouses had zero taxable investment income in the prior year, which would tend to imply a transfer of funds between spouses.
 Defined as individuals aged 65 and older.
 Bernheim (2002), p. 1,199.
 Throughout this paper, total income refers to the amount reported on an individual’s federal income tax return.
 OECD (2007), p. 35. HM Revenue and Customs (2007) and Joulfaian and Richardson (2001) reach a similar conclusion for tax-preferred savings accounts in the United Kingdom and the United States, respectively.
 While this section focuses on distributions for 2011, the profile of account holders in prior years is largely the same as in 2011.
 The average TFSA contribution amount for account holders with total incomes above $80,000 was about $4,340 compared to about $3,690 for account holders with total incomes below $80,000.
 About 248,000 individuals in 2011 made TFSA contributions that were greater than their incomes reported for tax purposes. The high level of contributions relative to income could represent a transfer of existing assets held by the individual into a TFSA or the transfer of funds between members of a household.
 In determining the household income of these individuals, the analysis is restricted to married and common-law couples (other potential members of the household are not considered).
 GIS benefits are reduced by 50 cents for each dollar of taxable income received by the pensioner and, in the case of couples, the pensioner’s spouse or common-law partner, other than Old Age Security benefits and the first $3,500 of employment income per person.
 Including individuals receiving the Allowance under the Old Age Security program.
 With respect to the type of savings individuals choose to shelter, it is expected that individuals would shift interest-bearing assets into TFSAs first for two reasons: (1) personal income tax rates on interest income are higher than those on dividends (the Dividend Tax Credit, which notionally recognizes taxes paid at the corporate level, reduces personal income tax payable on dividends); and (2) the transfer of interest-bearing assets is much less likely to activate a significant capital gains tax liability than the transfer of equity instruments (conversely, a capital loss arising from an in-kind contribution to a TFSA is disallowed).
 Using a more comprehensive definition of investment income which includes interest and other investment income, eligible Canadian dividends, and capital gains and losses, the percentage of tax filers reporting investment income or losses followed a similar pattern, falling from 38% in 2008 to 34% in 2009 to 31% in 2010, the last year for which such detailed data is available.
 It is necessary to disaggregate the investment income since interest, dividends and capital gains earned on non-registered assets are accorded different effective personal income tax rates.
 A weighted interest rate (based on interest rates available on TFSA saving deposits and the average yield of long-term Government of Canada bonds) is applied to saving deposits and fixed-income assets.
 The dividend yields of the S&P/TSX Composite index are applied to the stock of equity investments in order to estimate dividend income.