Longitudinal Immigration Database (IMDB) - Technical Report, 2015
Appendices

A) Links to key IMDB documents and web pages

Dictionaries (tax and immigration component):

Available to data users or upon request by contacting Statistics Canada by email at STATCAN.infostats-infostats.STATCAN@canada.ca

IMDB CANSIM tables:

http://www5.statcan.gc.ca/COR-COR/COR-COR/objList?lang=eng&srcObjType=SDDS&srcObjId=5057&tgtObjType=ARRAY

Historical IMDB:

http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getInstanceList&Id=7196

IMDB releases in The Daily:

http://www5.statcan.gc.ca/COR-COR/COR-COR/objList?lang=eng&srcObjType=SDDS&srcObjId=5057&tgtObjType=DAILYART

Analysis using the IMDB:

http://www5.statcan.gc.ca/COR-COR/COR-COR/objList?lang=eng&srcObjType=SDDS&srcObjId=5057&tgtObjType=STUDIES

The Consumer Price Index (62-001-X):

http://www5.statcan.gc.ca/olc-cel/olc.action?lang=en&ObjId=62-001-X&ObjType=2

Description of the annual Income Estimates for Census Families and Individuals (T1 Family File):

http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&SDDS=4105&lang=fr&db=imdb&adm=8&dis=2.

B) Coverage

The 2015 IMDB was used to produce these counts. Filers are linked immigrants who have filed a tax return at least once since 1982.

Table 15
Distribution of taxfilers and non-taxfilers by landing year
Table summary
This table displays the results of Distribution of taxfilers and non-taxfilers by landing year. The information is grouped by Landing year (appearing as row headers), Taxfilers, Non-taxfilers, Total, Immigrants, permanent resident, permanent resident | non-permanent resident and Deaths, calculated using number and percent units of measure (appearing as column headers).
Landing year TaxfilersTable 15 Note 1 Non-taxfilers Total
Immigrants PR PR | NPR Deaths Immigrants PR PR | NPR Deaths Immigrants PR PR | NPR Taxfilers
number percent
1980 121,400 116,500 4,900 17,900 21,700 21,200 600 1,500 143,100 137,700 5,500 84.8
1981 108,100 94,700 13,400 16,000 20,500 19,300 1,100 1,300 128,600 114,100 14,500 84.1
1982 103,800 87,700 16,100 14,200 17,300 16,200 1,100 1,200 121,100 103,900 17,200 85.7
1983 77,500 62,900 14,600 11,300 11,600 10,800 800 1,000 89,000 73,600 15,400 87.1
1984 77,800 61,200 16,600 10,600 10,200 9,400 800 900 88,000 70,600 17,400 88.4
1985 75,300 59,100 16,300 9,200 8,600 8,100 500 600 83,900 67,200 16,800 89.7
1986 89,200 65,400 23,800 9,500 9,600 9,000 600 500 98,800 74,400 24,400 90.3
1987 137,500 102,900 34,600 11,800 13,700 12,900 700 600 151,200 115,800 35,300 90.9
1988 145,300 127,700 17,600 11,300 15,400 14,500 900 500 160,800 142,300 18,500 90.4
1989 172,500 147,900 24,600 11,600 18,200 16,800 1,400 500 190,700 164,700 26,000 90.5
1990 194,400 160,400 34,000 12,200 21,100 19,400 1,700 500 215,400 179,700 35,700 90.3
1991 210,800 139,500 71,400 13,200 21,000 18,100 2,900 500 231,800 157,600 74,200 90.9
1992 231,200 146,700 84,500 13,300 22,700 19,200 3,500 400 253,900 165,800 88,100 91.1
1993 233,800 168,100 65,700 12,500 21,900 19,100 2,800 500 255,700 187,200 68,500 91.4
1994 202,000 163,100 38,800 10,100 21,600 19,900 1,700 400 223,600 183,100 40,500 90.3
1995 191,500 150,900 40,600 8,200 20,600 19,100 1,600 300 212,200 170,000 42,200 90.2
1996 201,700 159,200 42,500 7,000 23,700 22,000 1,700 300 225,400 181,200 44,200 89.5
1997 192,500 156,300 36,200 5,800 22,900 21,600 1,300 200 215,500 177,900 37,500 89.3
1998 158,700 126,400 32,300 4,200 15,000 13,800 1,200 190 173,700 140,200 33,500 91.4
1999 172,600 137,800 34,800 4,000 16,800 15,600 1,200 150 189,400 153,400 36,000 91.1
2000 205,200 165,300 39,900 4,100 21,600 20,200 1,400 180 226,700 185,400 41,300 90.5
2001 222,900 183,800 39,100 4,200 26,900 25,300 1,500 170 249,800 209,100 40,700 89.2
2002 199,700 166,500 33,200 3,700 28,600 27,200 1,300 170 228,200 193,700 34,500 87.5
2003 190,000 156,700 33,300 3,100 30,600 29,200 1,300 140 220,500 185,900 34,600 86.2
2004 199,300 156,200 43,100 2,500 36,000 34,100 1,900 130 235,300 190,300 45,000 84.7
2005 217,600 167,800 49,800 2,100 44,200 41,700 2,500 80 261,800 209,400 52,300 83.1
2006 208,300 154,200 54,100 2,100 42,800 40,100 2,700 110 251,100 194,200 56,900 83.0
2007 193,300 140,900 52,400 1,700 42,900 40,200 2,700 90 236,200 181,100 55,100 81.8
2008 198,200 143,700 54,500 1,500 48,400 45,800 2,600 90 246,600 189,500 57,100 80.4
2009 201,500 143,700 57,800 1,300 50,100 47,200 3,000 90 251,600 190,800 60,800 80.1
2010 216,500 156,000 60,500 1000 63,600 59,800 3,800 70 280,100 215,800 64,300 77.3
2011 189,600 135,300 54,300 700 58,500 54,400 4,100 40 248,100 189,700 58,500 76.4
2012 196,400 135,500 60,900 600 60,800 56,400 4,500 10 257,200 191,900 65,300 76.4
2013 195,100 130,500 64,600 500 63,400 58,500 4,900 10 258,500 189,000 69,500 75.5
2014 192,000 109,400 82,600 180 67,500 61,500 6,000 10 259,500 171,000 88,600 74.0
2015 185,300 102,500 82,800 30 85,700 78,500 7,200 0 271,000 181,000 90,000 68.4
Total 6,308,500 4,782,400 1,526,200 243,210 1,125,700 1,046,100 79,500 13,430 7,434,000 5,828,200 1,605,900 84.9
Table 16
Proportion of linked taxfilers by age group at landing, sex and admission decade
Table summary
This table displays the results of Proportion of linked taxfilers by age group at landing, sex and admission decade. The information is grouped by Sex and cohorts (appearing as row headers), Age at landing, 0 to 14, 15 to 24 , 25 to 34, 35 to 44, 45 to 64, 65 and older and Total, calculated using percent units of measure (appearing as column headers).
Sex and cohorts Age at landing
0 to 14 15 to 24 25 to 34 35 to 44 45 to 64 65 and older Total
percent
1980 to 1989 cohorts
Male 82.9 94.1 95.0 93.6 86.6 61.9 89.4
Female 82.2 92.0 93.2 92.8 83.1 59.9 87.2
Total 82.6 93.0 94.1 93.2 84.6 60.7 88.7
1990 to 1999 cohorts
Male 83.3 94.4 93.7 93.2 91.3 77.6 90.6
Female 82.0 94.6 94.3 93.7 90.3 77.0 90.5
Total 82.6 94.5 94.0 93.4 90.8 77.3 90.7
2000 to 2009 cohorts
Male 50.2 95.7 92.7 93.2 93.3 89.7 83.6
Female 49.2 96.0 94.5 94.6 94.1 88.6 85.4
Total 49.8 95.9 93.7 93.9 93.7 89.1 84.2
2010 to 2015 cohorts
Male 6.7 85.4 94.2 92.3 90.0 84.4 73.1
Female 6.8 87.8 94.8 93.7 90.3 83.4 76.1
Total 6.7 86.7 94.5 93.0 90.2 83.8 73.4

C) Previous analysis

Since its creation, the IMDB has been used to produce several analyses. The following is a summary of some Statistics Canada studies that have made use of the IMDB.

In recent years, several releases in The Daily have featured the IMDB. The subjects discussed include changes in the regional distribution of new immigrants to Canada, income and mobility of immigrants, immigrants in the hinterlands, and immigrants who leave Canada. These articles are accessible via the Statistics Canada website.

Papers using the IMDB have been published in the Perspectives on Labour and Income publication series (75-001-X) and the Analytical Studies Branch Paper Series. Among the topics covered were the income of immigrants who pursue postsecondary education in Canada, and the earnings advantage of landed immigrants who were previously temporary residents in Canada.

D) Best practices and tips for analysts

D.1 Programming tips

This section provides programming information for individuals who want to have a better understanding of the programming structure used to access data from IMDB files. Please note that individuals may conduct their own programming. There are two types of IMDB files—the yearly IMDB data files and the immigration data (for more details on IMDB files, refer to Section 3). IMDB tax variables are identified with a variable name that consists of three parts: (1) the acronym name as described in the IMDB tax data dictionary, (2) the aggregate level (I or F), and (3) the year (the four-digit year extension exists in most, but not all, cases).

Example: The interest and investment income at the individual level for 2014 would be named INVI_I2014.

Observations in the IMDB files are sorted according to a variable, IMDB_id (note that there is no year extension for this variable), which enables users to maintain a link across years. Data access takes place by means of the SAS programming language. A sample SAS program designed to access IMDB data is provided below. The samples below are created to perform the following task:

"retrieving the number of Social Assistance (SA) recipients for immigrants who landed between 2000 and 2005, living in Ontario between 2010 and 2012, and did not have any earnings appearing on their T4 slips by sex and year (2010 to 2012)"

Researchers who are new to the IMDB are encouraged to go through this sample SAS program. There are generally three components in the sample.

  1. Library set-up: The library assignments on the first two lines are the locations for the input files (first line) and the output files (the second line).
  2. Steps to generate a working dataset:
    1. The input files are stored in SAS format and can therefore be accessed with a SET or MERGE statement.
    2. This program is aimed at retrieving the number of Social Assistance (SA) recipients for immigrants who:
      1. landed at any time from 2000 to 2005
      2. lived in Ontario from 2010 to 2012
      3. did not have any earnings on their T4 slips
    3. And generate the number of SA recipients by sex and year (in this case, 2010 to 2012).
  3. The dataset used to produce the number of the SA recipients: The part, which starts with "proc freq," produces the numbers of interest as they are specified in the rest. At the end of the program, four tables are created from the output data file.

It is generally recommended that programs use the variables available in the PNRF rather than the yearly tax files for consistency. For example, the sample program uses the variable GENDER, a variable found in the PNRF, rather than SXCO_I&year, the variable found in the yearly IMDB_T1FF. In this program, only individuals who have filed every year from 2010 to 2012 are selected.

When programming in SAS, one should keep in mind the distinction between missing values and zeros in numeric fields. With SAS, most mathematical operations performed with missing values will return missing values. In IMDB, in years that an individual is present, numeric variables not relevant to that individual have a value of "0" (zero). For example, if a person without a spouse filed in 2010, the value for RRSPSI2010 (contributions to a spouse's RRSP) should be "0" (zero). If that individual did not file in 2010, the value will be missing.

Sample IMDB program

* Sample SAS program using the IMDB;

libname source1 '\\f8prod05\cic\1.Database'; * location of IMDB files ;

libname Out '\\f8prod05\cic\3.workarea'; * user's directory ;

* This sample program's objective is to use the IMDB to retrieve the number of Social Assistance (SA) recipients in Ontario that did not have any earnings appearing on their T4 slips, according to sex and year (in this case, 2010 to 2012). Data for provinces and earnings are from the yearly IMDB files whereas the sex variable is from the PNRF_2015. ;

* The first step is to create a datafile containing all the information that we need to produce our tables. This datafile will be called SAOnt and will be saved in the 'out' directory. The Longitudinal Identifier Number (IMDB_ID) is used to merge the annual IMDB datasets. ;

data out.SAOnt;
merge
source1.imdb_t1ff_2010(where=(prco_i2010 = 5 and outlier_ind2010 = 0) in=a keep=imdb_id prco_i2010 outlier_ind2010 saspyf2010 t4e__i2010)

source1.imdb_t1ff_2011(where=(prco_i2011 = 5 and outlier_ind2011 = 0) in=b keep= imdb_id prco_i2011 outlier_ind2011 saspyf2011 t4e__i2011)

source1.imdb_t1ff_2012(where=(prco_i2012 = 5 and outlier_ind2012 = 0) in=c keep= imdb_id prco_i2012 outlier_ind2012 saspyf2012 t4e__i2012)

source1.pnrf_2015(keep= imdb_id gender landing_year immigration_category);

by IMDB_id ;

If a and b and c and (landing_year>=2000 and landing_year<=2005);

*person must be taxfiler in all three years, not be flagged as an outlier, and must have landed between 2000 and 2005 (population of interest);

* We create a flag variable that identifies the SA recipients for each year. The result is three variables,
flag_sa2010, flag_sa2011 and flag_sa2012, taking a value of either 1 or 0.;
If (t4e__i2010=0 and saspyf2010>0) then flag_sa2010 = 1 ;
else flag_sa2010 = 0 ;
if (t4e__i2011=0 and saspyf2011>0) then flag_sa2011 = 1 ;
else flag_sa2011 = 0 ;
if (t4e__i2012=0 and saspyf2012>0) then flag_sa2012 = 1 ;
else flag_sa2012 = 0 ;
run;

* The SAS 'freq' procedure is used to produce our tables. We would also need to make sure that confidentiality guidelines standards are respected. ;

proc freq data = out.SAOnt;
tables immigration_category*flag_sa2010*flag_sa2011*flag_sa2012
gender*flag_sa2010*flag_sa2011*flag_sa2012 /missing ;
run ;
* End of the sample program;

D.2 Creating a cohort

Prior to starting an analysis, the cohort of interest needs to be defined. The cohort can be restricted by landing year, geography, or any other variable of interest (e.g., admission category or gender) according to the researcher's need. A clearly defined single cohort should be followed to allow comparability. For example, a researcher might be interested in women who landed in 2000 and who lived in a family that received social assistance in 2001 (Table 17). A study question regarding this cohort could be "What proportion of this cohort received social assistance in the following two years (2002 and 2003)?" It is worth noting that the Canada Revenue Agency (CRA) requires the spouse with the higher net income to report the social assistance payment. As a result, measurement on social assistance (SASPY_F), even for individuals, is best reported with the family-level information.

Table 17
Example—Women who landed in 2000 and received social assistance (SASPY_F) in 2001
Table summary
This table displays the results of Example—Women who landed in 2000 and received social assistance (SASPY_F) in 2001. The information is grouped by IMDB_ID (appearing as row headers), Landing year, Gender, SASPY_F2001, SASPY_F2002 and SASPY_F2003, calculated using dollars units of measure (appearing as column headers).
IMDB_ID Landing year Gender SASPY_F2001 SASPY_F2002 SASPY_F2003
dollars
IM583 2000 Female 20,500 19,000 14,000
IM145 2000 Female 3,000 0 0
IM548 2000 Female 11,500 13,800 0
IM798 2000 Female 16,000 18,000 8,000
IM961 2000 Female 10,000 0 0
IM967 2000 Female 9,500 0 0
IM110 2000 Female 5,000 2,000 1,000
IM125 2000 Female 1,000 0 200

D.3 Calculating retention rates

A key strength of the IMDB is the presence of geographic variables that allow for the study of mobility and retention. No other dataset contains a comparable level of detail on taxfilers annually, especially when it comes to smaller geographies. Having annual provincial, census division (CD), census metropolitan area (CMA), census agglomeration (CA), census subdivision level (CSD), and census tract level updates allows for a broad range of analyses.

Individual mobility trajectories can be studied simply by flagging changes in postal codes, and mobility trends can be calculated by studying relocations at specific levels of geography. For example, CSD-level mobility (year-to-year changes in CSD) and provincial mobility (year-to-year changes in province) significantly vary by a number of immigrant characteristics, such as age and admission class. These geographies are derived from the postal code (IMDB variable PSCO at the individual and family levels). The postal code is a six-character alphanumeric code that locates the point of delivery of mail addressed to post office customers in Canada. See Section 3.4.1 for a description of the geography variables.

In the example below (Table 18), the researcher is interested in mobility until 2002. IM798, IM961, IM967 and IM110 could be excluded from the mobility study because data (or files) are missing.

Table 18
Example—mobility until 2002 of immigrants who landed in 2000
Table summary
This table displays the results of Example—mobility until 2002 of immigrants who landed in 2000. The information is grouped by IMDB_ID (appearing as row headers), Landing year, Destination province, PRCO 2000, PRCO 2001 and PRCO 2002 (appearing as column headers).
IMDB_ID Landing year Destination province PRCO 2000 PRCO 2001 PRCO 2002
IM583 2000 B.C. B.C. B.C. B.C.
IM145 2000 Alta. Alta. Sask. Sask.
IM548 2000 Alta. Ont. Ont. Ont.
IM798 2000 Ont. Note ..: not available for a specific reference period Ont. Ont.
IM961 2000 N.B. N.B. N.B. Note ..: not available for a specific reference period
IM967 2000 Ont. Note ..: not available for a specific reference period Alta. Ont.
IM110 2000 Note ..: not available for a specific reference period Que. Note ..: not available for a specific reference period Que.

While mobility, at the individual level, is fairly straightforward, retention of immigrants in a jurisdiction can be calculated in several ways. How retention is calculated is an analytical decision based on the individual researcher's particular needs. The number of individuals retained is fairly straightforward to define—it is the number of individuals filing taxes in the jurisdiction of interest at a given time. A decision has to be made about what constitutes the initial landing cohort about which retention is calculated (the denominator in the retention rate).

The provincial rates reported in The Daily are defined as the proportion of immigrant taxfilers who reside in the province where they landed (defined as the province of intended destination) at a given time. For a given cohort (e.g., landing year) and a given tax year (or years since landing), the denominator is the number of taxfilers with the selected province of landing. The numerator is the number of taxfilers with the selected province of landing who are also residing in the province.

For example, using CANSIM table 054-0003 to compute retention rates three years after landing for the 2011 cohort, a researcher would choose all provinces of landing (i.e., the province of intended destination), all provinces of residence, landing year = 2011, and reference year = 2014. The table would look as follows:

Table 19
Province of residence in 2014 and province of landing, 2011 cohort
Table summary
This table displays the results of Province of residence in 2014 and province of landing, 2011 cohort. The information is grouped by Province of landing (appearing as row headers), Province of residence, Total province of residence, Newfoundland and Labrador, Prince Edward Island, Nova Scotia, New Brunswick, Quebec, Ontario, Manitoba, Saskatchewan, Alberta, British Columbia and Other residence, calculated using number of immigrants units of measure (appearing as column headers).
Province of landing Province of residence
Total province of residence Newfoundland and Labrador Prince Edward Island Nova Scotia New Brunswick Quebec Ontario Manitoba Saskatchewan Alberta British Columbia Other residence
number of immigrants
Total province of landing 174,740 405 330 1,365 880 31,505 70,590 9,695 6,120 26,965 26,390 500
Newfoundland and Labrador 515 325 0 5 0 5 75 5 0 60 30 0
Prince Edward Island 1,245 0 265 25 10 30 560 0 0 50 295 0
Nova Scotia 1,460 10 5 1,080 10 25 185 0 5 90 30 10
New Brunswick 1,340 0 10 35 750 55 275 0 10 80 120 0
Quebec 36,275 10 10 35 15 30,200 3,255 40 75 1,190 1,400 45
Ontario 69,135 35 25 115 70 875 63,145 275 335 2,815 1,325 115
Manitoba 11,190 0 0 15 0 55 645 9,170 80 825 380 10
Saskatchewan 6,360 0 0 0 0 20 295 45 5,370 445 165 10
Alberta 21,940 10 0 20 0 95 810 65 140 20,170 590 35
British Columbia 25,000 5 0 30 5 140 1,330 85 100 1,200 22,030 70
Other 280 0 0 0 0 0 15 0 0 35 20 200

Results for Nova Scotia shed some light on the matter. A total of 1,460 individuals landed in Nova Scotia in 2011 and filed taxes in 2014. Of those, 1,080 had Nova Scotia as their province of residence in 2014. Nova Scotia's three-year retention rate would be 1,080/1,460, or about 74%. The CANSIM table also provides information on secondary migrantsNote 11,365 individuals who landed in 2011 resided in Nova Scotia in 2014, of which 1,080 intended to land in Nova Scotia, and 285 had a destination province other than Nova Scotia.

The above definition of retention assumes that the number of taxfilers with the specific province of intended destination is the total population that can be retained in a year (i.e., if all 1,460 individuals who had intended to land in Nova Scotia had filed taxes there in 2014, the province would have 100% retention). This method does not take into account late or sporadic tax filing behaviour. While the total population in the 2014 tax year for Nova Scotia is 1,460, in the 2013 tax year 1,425 immigrants who intended to land in Nova Scotia filed taxes.

One alternative is a purely longitudinal approach, where a single landing cohort is selected (according to the province of intended destination, the province of initial tax filing, or both), and the retention rate is calculated as the proportion of this cohort that is still filing taxes in the province. When the province of initial tax filing is used to define the landing cohort, it is recommended that the first tax file occur in the year the immigrants were admitted (landing year = tax year), to exclude individuals who may have first arrived elsewhere and subsequently migrated to the region before filing taxes for the first time. A further restriction can be made if a researcher is interested in the population whose destination geography matches the geography of the first tax file.

Given that a portion of each annual cohort do not file taxes for their year of landing, it may be necessary to increase the population size for a region by defining the landing cohort as anyone who first filed taxes in the region within two years of landing (i.e., first_tax_year = landing_year or landing_year+1). Allowing individuals whose first tax filing occurred several years after landing to be part of a "landing cohort" is not recommended, as it is possible that they first landed elsewhere but did not file taxes. It is also a good idea to exclude intermittent filers from these analyses, as their place of residence is unknown in the years for which there is no tax data. Retention calculated this way will show a gradual decline in numbers; this decline is due to immigrants who stop filing, out-migration, and death.

If researchers are interested in secondary migrants to a region, this can be found by removing individuals in the defined landing cohort from the total number of immigrants filing taxes in the region at the time of interest. Again, however, these analyses should be restricted to individuals who first filed taxes within the same time period (year 0 or year 1) to avoid mistaking late-filers for in-migrants. If the landing cohort is restricted to immigrants whose destination geography matches the geography of first tax filing, a subsequent distinction should be made between secondary migrants who first filed elsewhere (and subsequently filed in the region of interest) and immigrants who first filed in the region of interest but were subsequently recruited by other jurisdictions (or information on their intended destination is missing altogether).

The following table presents an example of a longitudinal approach to provincial retention using fictitious data, with various definitions of the initial landing cohort.

Table 20
Number of immigrant tax filers within the specified population residing in British Columbia and associated retention rate, by years since landing
Table summary
This table displays the results of Number of immigrant tax filers within the specified population residing in British Columbia and associated retention rate, by years since landing. The information is grouped by Years since landing (appearing as row headers), Taxfilers who first filed taxes in British Columbia in year 0, Retention rate, Taxfilers who first filed taxes in British Columbia in year 0 or 1 and Taxfilers who first filed taxes in British Columbia in year 0 or 1 and province of intended destination was British Columbia, calculated using number and percent units of measure (appearing as column headers).
Years since landing Taxfilers who first filed taxes in B.C. in year 0 Retention rate Taxfilers who first filed taxes in B.C. in year 0 or 1 Retention rate Taxfilers who first filed taxes in B.C. in year 0 or 1 and province of intended destination was B.C. Retention rate
number percent number percent number percent
0 20,000 100 20,000 Note ...: not applicable 17,500 Note ...: not applicable
1 18,000 90 25,000 100 19,000 100
2 17,000 85 23,000 92 18,000 95
3 16,500 83 22,000 88 17,500 92

In the above example, retention in British Columbia can be calculated according to three definitions of the population, and the three-year retention rate varies per the definition adhered to. Importantly, all individuals in the sample filed taxes at each point in time.

Finally, analysts should use caution when studying low-level census geographies over a long period of time, as CA and CMA boundaries change and CSDs are dropped and added. If possible, analysts should run the Postal Code Conversion File (PCCF+) program to standardize postal codes to a constant census geography.

D.4 Calculating income trajectories over time

As is the case with retention, calculating year-to-year changes in wages, salaries and commissions earnings (or, for that matter, any economic variable) requires consecutive information. For example, if a researcher wants to compare the median wages, salaries and commissions earnings of the 2000 cohort of women aged 24 to 54, 1 year after landing and 5 years since landing (Table 21), records with missing T1FF files could be removed from the analysis. The decision to remove these records would be based on the desire to evaluate the cohort's median income versus the cohort filer's median income.

Table 21
Median wages, salaries and commissions earnings of the 2000 cohort of women aged 24 to 54, 1 year after landing and 5 years since landing
Table summary
This table displays the results of Median wages, salaries and commissions earnings of the 2000 cohort of women aged 24 to 54, 1 year after landing and 5 years since landing. The information is grouped by IMDB_ID (appearing as row headers), Landing year, Age at landing, Gender, Wages income 2001 and Wages income 2005, calculated using dollars units of measure (appearing as column headers).
IMDB_ID Landing year Age at landing Gender Wages income 2001 Wages income 2005
dollars
IM583 2000 34 Female 20,500 49,000
IM145 2000 53 Female Note ..: not available for a specific reference period 56,000
IM548 2000 29 Female 11,500 33,800
IM798 2000 31 Female 36,000 0
IM961 2000 42 Female 10,000 Note ..: not available for a specific reference period
IM967 2000 40 Female Note ..: not available for a specific reference period Note ..: not available for a specific reference period
IM110 2000 35 Female 0 59,000

Use caution when calculating the "first year in Canada" income as it might not represent a full year of taxation. For example, someone who landed in November of 2013 and filed taxes for 2013 would have only two months of income in 2013. A best practice is to use the first full year of income (landing year +1, see Table 19). One exception is pre-filers, those who filed taxes in Canada before landing and filed at landing year as well, are most likely reporting income for the entire year.

Over-time income should also be studied in constant dollars. Consequently, Consumer Price Index (CPI) adjustments should be made (Appendix D.7). This adjustment is made in the IMDB CANSIM tables.

D.5 Rounding data

Respecting the privacy of Canadians is important to Statistics Canada. Consequently, any tables produced from IMDB_TIFF files are subject to rounding. The purpose of rounding is to ensure that no small cells are released that may reveal information on specific individuals or small groups of individuals. In general, the macros will take an unrounded input dataset of various statistics (counts, means, medians, etc.) and output a rounded dataset.

The rounding rules are confidential, but the rounding macros are available to all researchers. Documentation describing how to use the macros is available. These macros are applied to the output tables of all researchers, to all external data requests, and to the released CANSIM tables.

D.6 Identifying outliers

The variable OUTLIER_IND was created to identify outliers within the T1FF (see Section 5.5). It should be used to remove outlier data from any calculation (e.g., mean, median, or regression) employing tax data. Outliers differ from one year to another, meaning that a person's data may be identified as an outlier for a given year but not for a subsequent year.

The following table gives the distribution of the outliers in the tax files for 1982 and subsequent years by type of resident for the 2015 IMDB. Less than 0.1% records were identified as outliers per tax year. The proportion of outliers increased from 1995 to 1996 as a result of updates to the outlier detection method applied to tax files for 1997 and subsequent taxation years.

Table 22
Distribution of outliers by tax year
Table summary
This table displays the results of Distribution of outliers by tax year. Permanent resident, Permanent resident with Non-permanent resident permit and Total, calculated using number and percent units of measure (appearing as column headers).
PR PR with NPR permit Total
number percent
1982 10 10 20 0.01
1983 50 10 60 0.02
1984 70 20 90 0.03
1985 30 0 40 0.01
1986 50 10 60 0.01
1987 60 10 70 0.01
1988 80 20 100 0.01
1989 70 20 90 0.01
1990 60 20 80 0.01
1991 70 30 100 0.01
1992 120 40 160 0.01
1993 150 70 220 0.01
1994 60 20 90 0.01
1995 150 80 240 0.01
1996 280 180 450 0.02
1997 340 230 570 0.03
1998 460 250 710 0.03
1999 380 230 610 0.03
2000 460 270 730 0.03
2001 450 260 700 0.03
2002 500 270 770 0.03
2003 440 240 670 0.02
2004 530 280 800 0.02
2005 550 310 850 0.02
2006 600 320 910 0.03
2007 580 340 920 0.02
2008 600 400 990 0.02
2009 720 380 1,100 0.03
2010 680 420 1,090 0.02
2011 860 490 1,350 0.03
2012 620 390 1,020 0.02
2013 780 470 1,250 0.03
2014 760 410 1,170 0.02
2015 750 390 1,140 0.02

D.7 Adjusting income for the Consumer Price Index (CPI)

In order to take into account the cost of living, all incomes should be adjusted to the Consumer Price Index (CPI) for Canada. "The Consumer Price Index (CPI) is an indicator of changes in consumer prices experienced by Canadians. It is obtained by comparing, over time, the cost of a fixed basket of goods and services purchased by consumers. Since the basket contains goods and services of unchanging or equivalent quantity and quality, the index reflects only pure price change."Note 2 The adjustment factors for 2015 are available in Table 23. To transform data to constant dollars of a specific year (base year), data users need to multiply the dollar values in all but the base year by a year-specific adjustment factor. To obtain the adjustment factors, data users need to divide the CPI of the base year by the CPI of the specific year. In Table 23, the base year is 2015.

Table 23
2015 Consumer price index adjustment factors
Table summary
This table displays the results of 2015 Consumer price index adjustment factors. The information is grouped by Year (appearing as row headers), 2015 consumer price index adjustment equals 126.6 divided by, calculated using number units of measure (appearing as column headers).
Year 2015 consumer price index adjustment equals 126.6 divided by
number
1982 54.9
1983 58.1
1984 60.6
1985 63.0
1986 65.6
1987 68.5
1988 71.2
1989 74.8
1990 78.4
1991 82.8
1992 84.0
1993 85.6
1994 85.7
1995 87.6
1996 88.9
1997 90.4
1998 91.3
1999 92.9
2000 95.4
2001 97.8
2002 100.0
2003 102.8
2004 104.7
2005 107.0
2006 109.1
2007 111.5
2008 114.1
2009 114.4
2010 116.5
2011 119.9
2012 121.7
2013 122.8
2014 125.2
2015 126.6

D.8 Calculating key income measures

The IMDB CANSIM tables contain several income measures. Table 22 describes which variables of the T1FF are included in their calculation.

Table 24
Description of the Longitudinal Immigration Database income measures
Table summary
This table displays the results of Description of the Longitudinal Immigration Database income measures. The information is grouped by Measure (appearing as row headers), Components and Formula (appearing as column headers).
Measure Components Formula
Wages, salaries and commissions income Earnings from T4 slips; other employment income T4E__i + OEI__i
Self-employment income
Since 1988 Self-employment income from business, profession, commission, farm, and fishing; limited partnership income SEI__i + LTPI_i
Before 1988 Self-employment income from business, profession, commission, farm, and fishing SEI__i
Investment income
1982 to 1987 Interest and investment income; dividends; capital gains/losses, net taxable INVi_i + XDIV_i + (CLKGLi * 2)
1988 and 1989 Interest and investment income; dividends; capital gains/losses, net taxable INVi_i + XDIV_i + (CLKGLi * 3/2)
1990 to 1999 Interest and investment income; dividends; capital gains/losses, net taxable INVi_i + XDIV_i + (CLKGLi * 4/3)
2000 Interest and investment income; dividends; capital gains/losses, net taxable INVi_i + XDIV_i + (CLKGLi * 100/64.58)
Since 2000 Interest and investment income; dividends; capital gains/losses, net taxable INVi_i + XDIV_i + (CLKGLi * 2)
Employment Insurance benefits Employment Insurance benefits EINS_i
Social welfare benefits Social welfare benefits (use family-level) SASPYf
Total income Sum of all measures described above

It is to be noted that all outliers are removed from these calculations (Outlier_ind=1), that the variable Province of Residence at the End of the Year (PRCO_) is used to identify the province, and that all incomes are adjusted according to the Consumer Price Index (CPI) of the year of the most recent T1FF available. "Mean with income" is the mean income of immigrant tax-filers with income of the given type. "Median with income" is the median income of immigrant tax-filers with income of the given type.

Notes

Date modified: