Components of income: explaining some differences

View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

Archived

This page has been archived on the Web.

Earnings / employment income
Investment income
Retirement income
Government transfers
Other income
Total income
Income tax
After-tax income

In this section income is broken down into its components, to better explain the differences in the estimates produced by the four programs considered in this document. The data that will be compared in this section are contained in Appendix C and represent standardised aggregate measures for the ten provinces¹. The territories have been excluded for purposes of comparability, since SLID does not include the territories in its survey². 

Earnings / employment income

Earnings / Employment income

Aggregate income, 2005 Standardised name	SLID	T1FF	Census	SNA (IEA)
	($,000,000)
Employment income	640,580	635,274	658,064	656,025
Wages and salaries	585,081	592,120	611,059	605,467
Net income from self-employment	55,499	44,935	47,005	50,558
Net income from farm SE	2,497	1,205	2,120	1,210
Net income from non-farm SE	53,002	43,730	44,885	49,348

Earnings or employment income is, by far, the major component of income. It is made up of wages and salaries and self-employment income. These will be addressed in turn.

Wages and salaries (reported on lines 101 and 104 of the T1 return) includes, in addition to regular pay, overtime, tips, commissions, bonuses and a number of other items.

The unadjusted figure for the SNA is considerably larger, primarily because the figure that is published in the Income and Expenditure Accounts (catalogue 13-001) is for wages and salaries and supplementary labour income (SLI). SLI includes employers' social contributions, most notably employer contributions to the CPP or QPP, workers' compensation plans, pension plans and other benefit plans as well as retirement allowances.

As noted earlier, for the SNA, contributions to pension plans are recorded as income, while the benefits received from these plans are considered a reduction in assets. The retirement allowances are expected to be counted as income in all three sources even though they are usually provided in a non-recurring way. This decision was mainly practical due to the non-availability of distinctions in the tax data used as principal source.

A second adjustment that must be made is to add royalties, which are included in wages and salaries for SLID, the Census and T1FF (as they are reported on line 104 of the T1 tax return) but are not included in the SNA where they are considered part of Interest, dividends and miscellaneous investment income. All adjustments made to the SNA data are summarized in Appendix B.

This brings the aggregate SNA figure much closer to those of the other three programs. Wages and salaries are somewhat higher for the Census than for SLID and T1FF, as can be seen above. It is even more dramatic when considering that because of the Census net under-coverage, the population reported is 2.8% lower than the official estimates. There is no information to help assess the income levels of the under-covered (though if they were similar, one might expect the aggregate 2.8% lower).

Possible explanations for these differences might include misclassification of the income component - income that is truly from self-employment might be reported as wages and salaries; miscapture of paper reported amounts and reporting amounts not formally documented as income with CRA. The estimates published by SLID and T1FF are very similar. The small difference (less than 1%) could be attributed to sampling variability or to the fact that SLID excludes the income of certain groups of persons, as indicated in Appendix A (e.g. those living on Indian reserves, residents of institutions and the military living in barracks).

Table 3 presents distributional statistics for data sources which can provide disaggregated data. These highlight that using the aggregate measures for comparisons may mask differences and similarities between estimates of different points of the distributions. Thus, the comparison of aggregate data is not sufficient to identify dissimilarities and reconcile the data sources for all analysis projects³.

Table 3 Wages and Salaries Distribution Statistics by Source for Canada (Provinces only), 2005

Income from self-employment is the second component of earnings.

Non-farm self-employment income represents the gross receipts of non-farm unincorporated businesses and professional practices, less operating expenses such as wages, rents and depreciation. Once again, the estimate from the Income and Expenditure Accounts of the SNA is considerably higher. This is because, as already noted, the SNA considers persons who own their home to be self-employed, and attributes to them an imputed income equivalent to the rental income they would receive if they rented their dwelling. SLID, T1FF and the Census do not include such a value in their income measures. As well, the SNA includes other rental income with self-employment income; SLID and the Census consider it investment income and T1FF considers it other income (see reconciliations of specific sources in Appendix B). The adjustments that must be made to the non-farm self employment income of the SNA are reported in Appendix B.

Again, this minimizes the difference between this estimate and those of the other three programs. SLID aggregate self-employment remains higher than the other adjusted sources - the calibration of SLID into income classes may possibly account for some of the apparent over-coverage of the self-employed (they are pooled with the under-represented non-wage earners). This area requires further study.

Farm self-employment income represents the gross receipts from farm sales less depreciation and the cost of operation. Gross receipts includes cash advances, dividends from co-operatives, rebates and farm support payments. Income in-kind, such as the value of agricultural products produced and consumed on the farm, are included in the SNA estimate but excluded from the estimates of the other programs. This in-kind income amounted to a relatively small proportion of total net farm income in 2005 ($126 million⁴).

The table below provides data from Agricultural Division's publication Net Farm Income.⁵ The figure that can be used to compare with the estimates from SLID, T1FF and the Census is $1,571,127,000 (calculated by deducting income-in-kind (126,261,000) from realized net income (1,697,388,000).

Components of net farm income, 2005

Income components	2005
	($000)
Cash receipts, total	36,683,856
Less: Operating expenses after rebates	30,497,956
Equals: Net cash income	6,185,900
Plus:   Income-in-kind*	126,261
Less: Depreciation charges	4,614,773
Equals: Realized net income	1,697,388
Plus:  Value of inventory change*	589,012
Net income, total	2,286,400
Source: Agriculture Division, Net Farm Income, Catalogue 21-010, Nov. 2007 Note:*Excluded from SLID, Census, T1FF

This figure compares quite closely with that of T1FF but is considerably lower than the estimate generated by SLID. Sampling variability could be an explanation for this. The farming population is relatively small and therefore the sample of those with farm income in SLID would be correspondingly small. For this reason, the coefficient of variation for this estimate is quite high (16% or higher) and the published figure is accompanied by a note indicating that the estimate should be used with caution. It cannot be said that it is significantly different from the Census but both estimates are still higher than T1FF and SNA data.

Investment income

Investment income

Aggregate income of recipients, 2005	SLID	T1FF	Census	SNA (IEA)
	($,000,000)
Investment income	29,012	38,883	37,755	39,812

For SLID, T1FF and the Census, the concept of investment income is very similar: it includes dividends, interest from such things as deposits, savings accounts, bonds, debentures, t-bills and loans, and other miscellaneous forms of investment income⁶. Rental income is included for SLID and the Census but not for T1FF and the SNA. The estimates of aggregate investment income for the Census and T1FF are very similar once adjusted for net rental income, although it is about 20% lower for SLID. This may partly be due to sampling variability and weight editing to reduce the influence of outliers but may also point to difficulties reaching the very high income people in a survey context.

As can be seen in the table above, the estimate produced by the SNA is significantly higher than for the other three programs. This is because, as previously noted, it includes the interest accrued during the year in the non-profit institutions serving households (mainly employer pension plans, RRSPs and life-insurance companies), as these investment earnings are considered income at the time they are generated. For SLID, T1FF and the Census these amounts would not be regarded as income until the benefit from these plans is received. The SNA also includes in investment income the interest on investments made by other "associations of individuals". When these amounts are deducted from the SNA estimate of investment income the result is very similar to the estimate produced by T1FF and the Census. Work is underway in the SNA to develop a pension satellite account which would show the flows in and out of these institutions⁷.

Retirement income

Standard table symbols

Retirement income

Aggregate income of recipients, 2005	SLID	T1FF	Census	SNA (IEA)
	($,000,000)
Retirement income	58,993	56,465	56,906	.
. not available for any reference period

Retirement pensions include all regular income received as the result of having been a member of a pension plan of one or more employers or of contributing to a Registered Retirement Savings Plan (RRSP). This income may be received in various forms: as an annuity, superannuation or Registered Retirement Income Fund (RRIF).   Lump-sum benefits are not included, although, for data obtained from tax returns, RRSP income reported on line 129 of the tax return is included, but only if the individual is 65 or older⁸.  Pensions paid to widows or children of deceased pensioners would also be included here. 

Processing for the 2006 Census was slightly different in one respect than that for SLID and T1FF: an additional procedure sought to remove from income any lump-sum transfers to RRSPs. These transfers are identified on the tax form and may come from RRSP withdrawals, pension income or the other income fields on the tax return. These are conceptually not income since they represent shifts from one savings vehicle to another. This adjustment will decrease the aggregate Retirement income as it reduces the amounts reported in pension income and for RRSP withdrawals (only those of those aged 65 years or older are counted as income).

The concept of retirement income is essentially the same for SLID, T1FF and the Census and the estimates also are essentially the same though the Census is slightly lower than the other two sources which might be partially explained by the modified processing techniques. As noted earlier, the SNA does not consider pensions to be income at the time they are paid as benefits, but rather at the time the contributions are made and the investment earnings accrue. It is thus quite difficult to arrive at comparable or reconcilable estimates.

Government transfers

Standard table symbols

Government transfers

Aggregate income of recipients, 2005	SLID	T1FF	Census	SNA (IEA)	Admin. data9
	($,000,000)
Government transfers	96,633	97,969	95,591	110,394	..
Child benefits	10,153	11,204	11,278	11,664	..
Federal child benefits	7,872	9,129	..	9,129	9,129
Provincial child benefits	2,280	2,075	..	2,371	2,371
OAS and GIS	26,558	27,152	25,927	25,996	29,052
CPP and QPP benefits	28,705	30,144	29,018	29,261	31,614
Employment Insurance benefits	12,688	12,134	11,978	12,616	13,398
Other government transfers	20,402	17,334	17,390	30,858	..
Goods and Services Tax (GST) and     Harmonized Sales Tax (HST) credits	3,154	3,051	..	3,462	..
Workers compensation benefits	5,001	4,559	..	5,201	4,970
Social assistance	8,616	8,076	..	13,313	9,163
Provincial and territorial tax credits	1,758	1,647	..	..	..
Military pensions and allowances     and miscellaneous transfers	1,873		..	..	..

Government transfers include all transfer payments from federal, provincial or municipal governments programs intended to provide income support to certain groups, such as seniors, families and those injured on the job. Specific major programs are listed in the table above.  For the SNA, it also includes transfers to "associations of individuals", for example, to benevolent associations or Aboriginal associations and also includes scholarships and research grants¹⁰, which are included in other income or employment income for the other three programs. For that reason it is difficult to compare the estimate from the SNA with that of SLID, T1FF and the Census. It is not surprising, however, that it is considerably higher than the other estimates.

Included in the Government transfers table above, in addition to the estimates of the four programs examined in this report, are data from administrative sources. These data are not readily available for all transfer income components but several can be produced as special tabulations of various tax slips obtained from CRA. These slips were generally produced by the payer and sent to the income recipient to assist when completing their tax return and a copy was made for CRA as well.

For other small governmental programs, it may not be possible to obtain micro-record data (especially if they are non-taxable) and they are not practical to measure or impute precisely in a survey context. SNA however, may include them in their government accounts as the global amount of each program may be more easily known.

Table 4 Sources of Administrative Data for Government Transfer Components

The administrative amounts usually have full coverage of amounts paid during the year (though for social assistance benefits that might be disputed) and as such are closer to the concept of the SNA because of the coverage limitations of the other sources with respect to deaths and institutional or collective populations. The T1FF is less affected as it re-assigns the income of the deceased to a surviving spouse if there is one and it is thought to have relatively good coverage of the institutional population.

The SNA uses most of these administrative sources to generate their estimates, which are consequently very close. The slightly lower estimates produced by T1FF may be due to non- or late filers. One would expect the estimates produced by SLID and the Census to be more comparable to those of T1FF, as tax data is the primary source for all three programs. This is indeed the case, although the estimates from SLID and the Census are somewhat lower than T1FF's (usually by less than 5%). For several programs catering mainly to older adults (such as OAS / GIS / Allowance program (usually with OAS and GIS) and CPP / QPP), the main discrepancy could be due to income recipients deceased during the year (or before Census day for the Census) and to persons living in institutional settings such as retirement homes or nursing homes which are generally excluded from the SLID population and the Census long form population. Some processing options in SLID and Census also compute entitlements and assign them as if the take-up rate was 100%.

Additional discrepancy at the conceptual level exists because OAS and EI have specific means-tested repayments listed on the tax return. It is possible to include as income only the part that will not be repaid at tax time (as Census and SLID do) or to use the full amount paid and increase the tax paid by the amount of these social benefits repayments (as T1FF and SNA do). This affects the individual components, total government transfers and total income. Regardless of the accounting chosen for a specific statistical program, the after-tax income concepts are identical for all. For EI benefits, the amounts repaid at tax time correspond to 1% of total benefits paid. For OAS repayments, this corresponds to 3% of benefits paid.

With respect to child benefits, the published estimate for T1FF is lower than that for SLID and the Census because T1FF tables include benefits from provincial and territorial child benefit programs in another category with provincial/territorial tax credits. T1FF's published estimate of provincial/territorial tax credits/benefits is therefore correspondingly higher than SLID's. Once adjusted, the SLID data seem almost 10% lower than estimates from other sources. Most of the differences come from the Federal child benefits component.

Under other government transfers, SLID and T1FF both generate estimates for specific sub-components (e.g., worker's compensation, social assistance and tax credits) using tax data or other administrative sources and, for those who do not provide permission to use their tax returns, SLID specifically asks respondents for the amount of income they receive from each of these programs. The Census, on the other hand, asks one question on other government transfers globally, providing examples of programs to consider; it is therefore possible that there is some response error, because respondents may not recall all of this income.

Some additional variations in published tables may exist because of publication practices. For confidentiality purposes, T1FF removes income from their tables for atypical circumstances even in the most aggregated, Canada-level data. For example, people without children at year end are assumed not to have received child benefits though that situation is possible on occasion if a child has died or custody has changed during the year. The dataset retains the original values so custom tabulations may take advantage of the atypical data for special analysis.

Other income

Standard table symbols

Other income

Aggregate income of recipients, 2005	SLID	T1FF	Census	SNA (IEA)
	($,000,000)
Other income	19,188	16,908	15,817	..

Other income is a residual category for the surveys including very different components and this causes some classification and comparability issues across the income statistics programs. For example, Other income in the T1FF estimates includes net rental income and income from limited partnerships. These are classified as Investment income in other sources. These two components can be measured and re-assigned but other discrepancies are more difficult to reconcile. In tax data, a variety of elements are reported on line 130 of the tax form under the Other income heading and these may include with market income some government transfers such as large scholarships. Another example is that transfers to an RRSP can sometimes appear here. Census procedures screen out the amounts recognised as transfers to a registered plan.

T1FF procedures identify non-filing spouses mentioned on their partner's return. The net income may be reported on the filing spouse's return. If no T4 slip was filed by the employer, this income is then usually classified quite arbitrarily into Other income¹¹. In 2005, 196 million dollars were classified in this way.

A proportion of the SLID figure includes income amounts reported in a catch-all category with respondents' description of the component (1,546 million dollars in 2005). In 2005, the extra component fields were not coded and the amounts were classified into Other income.

These differences may help explain why the Census estimate is less than the other two but may not fully explain. One singularity of the T1FF as opposed to Census and SLID is the treatment of alimony and child support amounts received. The T1FF focuses on the taxable portion of support payments received whereas Census and SLID are trying to measure the total transfer.

The amounts in this component, for the SNA, are mostly considered transactions between actors within the Persons and unincorporated businesses sector so these would not be considered personal income for SNA as they do not come from other sectors of the economy.

Total income

Standard table symbols

Total income

Aggregate income of recipients, 2005	SLID	T1FF	Census	SNA (IEA)
	($,000,000)
Total income	844,406	847,982	864,163	..

The difference between the lowest and the highest measure of aggregate total income from SLID, Census and T1FF is approximately 2.4%. The Census estimate is somewhat higher than either SLID's or T1FF's; some of the reasons have been examined when considering individual components (such as earnings).

For SNA, the comparable figure is not computed because of the conceptual differences related to the issue of how to account for the transfers between different members of the Persons and unincorporated business sector.

Income tax

Standard table symbols

Income tax

Aggregate income of recipients, 2005	SLID	T1FF	Census	SNA (IEA)
	($,000,000)
Income tax	145,294	152,974	153,160	163,936
Federal income tax	88,943	93,375	..	..
Provincial income tax	56,352	59,599	..	..

Income tax is not itself a component of income but it is useful for assessing the money available to households for current consumption and saving. The concepts of income tax are broadly similar across programs. The reference period is one crucial difference.

The concept used for the three micro-economic programs (SLID, Census and T1FF) is the income tax paid on the income of the reference year whether or not it was paid during the calendar year or at tax filing time (usually the following spring). Any amounts paid in April of the reference year for the previous year's balance of liabilities are not considered.

For practical and quality reasons, the SNA reference period is different, it is on a cash basis instead of using an accrual method referring to the period of the economic activity: all amounts paid during the year as income tax are counted whether they are for past, current or future liabilities net of any over-payments for prior years. This generates inconsistencies across programs and renders reconciliations with SNA data more difficult. The inconsistencies would be greater in periods when income tax rates or economic activity is more variable.

In Quebec, Ontario and Alberta some amounts are collected with the tax return but are labelled separately as Health taxes. These are included in the income tax concept. Other provinces may have a separate contribution program to their health insurance schemes but if they are not collected on the income tax returns, they are not considered income tax in these frameworks. As mentioned earlier, the refund of social programs (such as Employment Insurance Benefits and Old Age Security pensions) is considered by SLID and Census as reductions in program benefits but for T1FF and SNA they are considered included as income and an equivalent tax.

The full year concept is relatively difficult to obtain accurately from respondents if asked directly on a questionnaire (they often report refunds or final payment amounts only). This therefore makes after-tax income a challenge to measure for programs that are exclusively respondent-based. Since all programs considered here are now based on administrative data or a mixed mode approach, this problem is no longer as clearly observed in survey data but a portion of the sample is still respondent-based.

During processing, the taxes liable to the Quebec provincial government must be imputed. This calculation is complex and certain approximations are required as Statistics Canada does not have access in the federal file to all the inputs used by the Ministère du revenu du Québec and certain deductions are exclusively available on the provincial returns. This problem does not arise when computing only an aggregate measure such as for SNA as reports on aggregate income tax revenue are available from the provincial authority.

In addition, Census does not offer disaggregated federal and provincial taxes for 2005 and SNA requires consulting separate accounts to split the aggregated figure.

One remaining internal consistency issue common to all programs relates to the desire to compare income tax to Total income. However, the total income concept we have developed may include non-taxable elements (such as child benefits, social assistance, etc.) and not necessarily all taxable components as considered by CRA (such as capital gains and RRSP withdrawals). Therefore, the computation of taxation rates is subject to some discussion.

After-tax income

Standard table symbols

After-tax income

Aggregate income of recipients, 2005	SLID	T1FF	Census	SNA (IEA)
	($,000,000)
After-tax income	699,112	695,008	711,003	..

By construction, with total income and income taxes paid, SLID, T1FF and Census programs compute a measure of after-tax income. Across these three programs, the concept of after-tax is identical. However, while the SLID and T1FF agree quite closely (less than 1% difference for the aggregate), the Census is substantially higher. This follows from the higher amounts seen for total income and might be related to coverage issues in each of the statistical programs. Since total income is not equivalent in the SNA, the after-tax income that could be computed would not be conceptually equivalent either.

After-tax income proves useful to analyse the economic means available (usually to the family or household) after provision of the government transfers and payment of taxes. Other special-purpose measures also exist on these databases which differ from after-tax or total income. For example, disposable income is after-tax income from which spending amounts considered mandatory are removed (contributions to public and private pensions, child care in order to work, out-of-pocket medical expenses, spousal support,… ). On T1FF, an alternative version of total income that corresponds to line 150 on the tax return permits analysis according to a tax-based conceptual frame-work.

 

---

Notes

1. The unadjusted data as published and adjustments are available in Appendix B tables.
2. The Canada-wide measures are reported individually in Appendix B for the sources that permit it.
3. The distributional statistics and semi-aggregated statistics will be examined in a companion paper to be published later.
4. Net Farm Income , Statistics Canada Catalogue 21-010, November 2007, page 13.
5. Net Farm Income , Statistics Canada Catalogue 21-010, November 2007.
6. Investment income does not include capital gains, as noted in section 1.
7. "Preliminary results of the Pension Satellite Account, 1990 to 2007" in the Latest Developments in the Canadian Economic Accounts (13-605-X) shows a time series of pension assets.
8. This is because RRSP income as reported on line 129 could be either a withdrawal or an annuity. If the person is less than 65 the amount is assumed to be a withdrawal, and is therefore not considered income, as it would not be received on a regular basis.
9. The administrative data have not been adjusted conceptually and the sources are presented in Table 4.
10. For the non-SNA programs, scholarships and bursaries are included in the component Other income; net research grants are included under wages and salaries.
11. One exception is when the non-filing spouse is thought to be over 65 years of age. In this case, any amount up to basic OAS ($5,706 in 2005) is classified in the OAS / GIS field and the remainder, if any, would be sent to Other income. The net income amount reported on the spouse's return is not adjusted to include any possible deductions for RRSP or pension contributions, union dues, admissible child care expenses, etc. before counting it.