# Analytical Studies Branch Research Paper Series Immigrant Entrepreneurs as Job Creators: The Case of Canadian Private Incorporated Companies

by Garnett Picot and Anne-Marie Rollin
11F0019M No. 423
Release date: April 24, 2019

## Abstract

This paper focuses on job creation and destruction by private incorporated companies with immigrant and Canadian-born owners, and uses data covering the 2003-to-2013 period. The unadjusted (raw) data indicated that average annual net job growth per firm was higher among immigrant-owned firms than among firms with Canadian-born owners, as was the likelihood of being a high-growth firm. Regression analysis revealed that these differences were largely because immigrant-owned firms were younger on average, and younger firms create jobs at a higher rate. Moreover, immigrant-owned firms accounted for a disproportionate share of entering firms, which play a significant role in job creation. Because of the constant inflow of new immigrants, immigration led to the creation of a large number of new young firms. Through this process, immigrant-owned firms contributed disproportionately to net job creation over the period studied.

## Executive summary

Using data from Statistics Canada’s Canadian Employer–Employee Dynamics Database (CEEDD), this paper has three objectives: (1) determining how the number of jobs created or destroyed by immigrant-owned private incorporated companies compared with that of firms with Canadian-born owners, (2) determining whether immigrant-owned firms were more likely than firms with Canadian-born owners to be high-growth firms or rapidly shrinking firms, and (3) determining which immigrant characteristics were associated with a higher likelihood of immigrant-owned firms being high-growth firms or rapidly shrinking firms. The data covered the 2003-to-2013 period. Only immigrants who have entered Canada since 1980 could be identified as immigrants in the study; they accounted for roughly three-quarters of all immigrants aged 18 to 69 in Canada in 2011.

This paper addresses gross job creation (jobs created by expanding continuing firms and entering firms), gross job destruction (jobs terminated by contracting continuing firms and exiting firms), and net job change (the difference between gross job creation and gross job destruction).

With regard to gross job creation and gross job destruction, the results revealed that private incorporated immigrant-owned firms were much more likely than firms with Canadian-born owners to be job creators than job destroyers. As a result, in terms of net job change, immigrant-owned firms held the advantage in the unadjusted (raw) data, registering a higher average annual net job creation per firm than firms with Canadian-born owners. For example, firms owned by immigrants who entered Canada since 1980—the population of immigrants available in the dataset used for this study—accounted for 25% of the net jobs created in the private incorporated sector over the 11-year period, while representing 17% of the firms studied.

Entering firms played a larger role in job creation among the immigrant-owned firm population than among firms with Canadian-born owners. The difference in job creation by entering firms and job loss by exiting firms led to one-half of all net job creation among immigrant-owned firms, and one-third of all net job creation among firms with Canadian-born owners. Finally, immigrant-owned firms in the sample accounted for a disproportionate share of entering firms compared with firms with Canadian-born owners.

The characteristics of immigrant-owned firms and firms with Canadian-born owners differed significantly. Notably, immigrant-owned firms were younger, and younger firms have higher job creation rates, which can affect job creation and destruction rates. After the results were adjusted for differences in these and other characteristics, the difference in average annual net job creation between immigrant-owned firms and firms with Canadian-born owners disappeared and, if anything, immigrant-owned firms had lower average annual net job creation rates.

The results were similar with respect to a firm’s likelihood of being a high-growth firm (over 20% annual growth in employment). The unadjusted results (raw data) suggested that immigrant-owned firms were significantly more likely (1.3 times as likely) than firms with Canadian-born owners to be high-growth firms. There was little difference in the likelihood of being a rapidly shrinking firm (below -20% annual growth in employment). When the results were adjusted for differences in the characteristics of the firm, the difference in the probability of being a high-growth or rapidly shrinking firm was reduced by 70%.

Immigrants are a very heterogeneous group. An immigrant owner’s educational attainment, source region or immigrant class (refugee, economic or family class) may affect the job creation rate of the immigrant-owned firm. However, the adjusted results suggested that the characteristics of the immigrant owners had relatively little effect on the likelihood of an immigrant-owned firm being a high-growth or rapidly shrinking firm.

How can these results be interpreted? Based on the raw data, immigrant-owned firms had a higher level of net job creation per firm, and were more likely to be high-growth firms than those with Canadian-born owners. But most or all of this gap was due to differences in the characteristics of the firms. Most notably, immigrant-owned firms were younger, and younger firms are more dynamic regarding job creation.

Immigrant-owned firms were younger precisely because they were owned by immigrants. A constant inflow of immigrants resulted in a larger share of young firms than would have been the case without immigration, as newly arrived cohorts of immigrants started new private incorporated companies. There was a “years since immigration” effect. Only after immigrant owners had been in Canada for over 30 years did their share of young firms (4 years old or younger) approach that of Canadian-born owners. Therefore, the proportion of young firms was higher among immigrant-owned firms than among firms with Canadian-born owners. This statement is supported in this paper both for firms owned by immigrants who have entered Canada since 1980 and for firms owned by all immigrants in Canada, including those who entered Canada prior to 1980. In 2012, approximately 48% of private incorporated firms owned by all immigrants were young, compared with 29% of those with Canadian-born owners. Thus, while the controls for differences in characteristics eliminated much or all of the job creation advantage immigrant-owned firms held over firms with Canadian-born owners, that was not the end of the story. Immigration increased the job creation dynamism of the private incorporated company sector by generating a large number of entering and young companies, which led to immigrant-owned firms’ disproportionate contribution to job creation in the private incorporated company sector.

## 1 Introduction

This paper focuses on job creation and destruction among private incorporated companies in Canada over the 2003-to-2013 period. Using the Canadian Employer–Employee Dynamics Database (CEEDD), the paper explores whether and why the employment dynamics of immigrant-owned firms differ from those of companies with Canadian-born owners. It addresses both the level of job creation and destruction, and the growth rate of private incorporated companies.

To achieve its goal, this paper merged two different research topics: the economic contribution of immigrant entrepreneurs, and the job creation by firms in the private sector. Immigrant entrepreneurship and its contribution to the host economy has been the topic of considerable recent research (see OECD [2010] and Fairlie and Lofstrom [2013] for reviews). In countries with high immigration rates (Canada, Australia, the United Kingdom and the United States), immigrants are seen as entrepreneurial—their tendency to be business owners surpasses that of the native-born population (Schuetze and Antecol 2007; Fairlie and Lofstrom 2013).

In their review of immigrant entrepreneurship, Fairlie and Lofstrom (2013, p. 38) concluded that “there is little evidence in the literature on how much immigrant-owned businesses contribute to job growth. Although data exists [sic] on employment among immigrant-owned businesses no data are available showing the dynamics of employment among these firms.” Recently, American studies looked at immigrant firm formation and employment growth. Presenting what they described as “tentative findings,” Kerr and Kerr (2016) concluded that, over the medium term, new immigrant-owned firms were more likely to survive and modestly more likely to experience employment growth than comparable firms founded by the American-born population. Anderson (2016) studied privately held American start-up companies valued at \$1 billion or more and noticed that about half of them had an immigrant founder. This provided circumstantial evidence that immigrants contributed disproportionately to the creation of new, fast-growing companies in the United States.

No Canadian research on the employment dynamics of immigrant-owned firms and firms with Canadian-born owners was found.Note  Whether immigrant-owned firms experience faster or slower employment growth is an unanswered question. The results from other countries may or may not apply to Canada. Immigrants to Canada differ in important ways from those entering the United States and Europe. They are typically more highly educated (Ostrovsky, Picot and Leung 2018) and more likely to be economic immigrants. This could influence both start-up rates and employment dynamics.

The literature on business job creation and destruction was another important resource for this study. There is a significant history that addresses the many issues encountered when attempting to assess job creation and destruction at the firm level (see Davis, Haltiwanger and Schuh 1996). A key finding in this literature was that firm size was a crucial variable when explaining job creation and destruction, but whether it was positively or negatively related to job growth depended on the data used (see Audretsch et al. [2004] and Geroski [2005] for a thorough literature review). A more recent result suggested that firm age also mattered, and its omission in earlier job growth studies partly explains the mixed results obtained for firm size. Young firms were key drivers of job creation (Haltiwanger, Jarmin and Miranda 2013).

For Canada, recent research examined whether the growth rate of Canadian employers varied between firms of different sizes and ages across the entire distribution (Dixon and Rollin 2012) and in the tails of the distribution (Dixon and Rollin 2014; Dixon, Petrunia and Rollin 2018; Decker et al. 2015). Younger firms of any size were found to be more likely to post high annual growth rates than older firms. The contribution of high-growth and rapidly shrinking firms to job creation and destruction in Canada was also evaluated. This study builds on the approach used by Dixon and Rollin (2012, 2014) and Dixon, Petrunia and Rollin (2018). However, it differs in that two measures of job change are employed (change in levels and growth rate), and the focus is on the difference between immigrant and Canadian-born owners.

This paper focuses on job creation in Canadian companies with immigrant and Canadian-born owners. There are three possible types of ownership structures: unincorporated businesses, privately owned incorporated companies and publicly owned incorporated companies. This paper focuses only on private incorporated companies. Unincorporated businesses (i.e., the unincorporated self-employed) were excluded from this study because very few (about 3% for immigrant-owned firms) have employees beyond the owner (Green et al. 2016). Publicly traded companies were also excluded because ownership is typically widely dispersed, and the owners are not necessarily involved in the day-to-day operation of the company. This paper focuses on Canadian-controlled incorporated companies in the private sector, which accounted for about 55% of private sector employment between 2003 and 2013. Understanding long-term trends is important. Trends such as those observed over the study period change slowly over time, and are likely applicable to more recent years.

## 2 Data and methods

### 2.1 Data source

The data come from the CEEDD, which consists of a number of components, including T2 corporate tax data, T1 individual tax data, T4 records of remuneration issued by employers, the Longitudinal Employment Analysis Program (LEAP) and the Longitudinal Immigration Database—a longitudinal file of all immigrants who have entered Canada since 1980. In 2011, roughly 75% of all immigrants in Canada aged 18 to 69 had entered Canada since 1980 according to the National Household Survey (Green et al. 2016). This dataset includes information on employees, firms, business owners and immigrants. See the study by Green et al. (2016) for a detailed description of the CEEDD.

The key to the research on immigrant entrepreneurship is the ability to determine the immigration status of each business owner. Until the recent creation of CEEDD data, this was not possible in Canada. If a firm had at least one immigrant owner, it was considered to be an immigrant-owned firm in this study. The main results were replicated using an alternative definition, where all owners must be immigrants for the firm to be considered immigrant-owned. The results based on the two different definitions were substantially the same (results available from the authors).

The immigration status of the owners could not be established for about 15% of private incorporated companies (Green et al. 2016). To correct for this underreporting, weights were applied to each record to adjust for non-reporting. These weights were estimated and applied at the level of province, firm age, firm size and industry cells.

The firm’s employment in any given year was obtained from the LEAP database. This employment estimate was based on the firm’s total payroll. A conversion factor derived from the Survey of Employment, Payrolls and Hours was used to convert the payroll to an employment estimate (Lafrance and Leung 2009). This estimate is referred to as an average labour unit (ALU).

### 2.2 Methods

The central variable in this research is the change in a firm’s employment—either change in employment levels or employment growth rate—between consecutive years from 2002/2003 to 2012/2013. The change in employment is simply the change in employment level between $t$ and $t-1$ .

To calculate the growth rate of firm $i$ from year $t-1$ to year $t$ , the average employment of the firm over these two years was used in the denominator:

$E i t − E i t − 1 ( E i t + E i t − 1 ) / 2 , MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaaSaaaeaaca WGfbWaaSbaaSqaaiaadMgacaWG0baabeaakiabgkHiTiaadweadaWg aaWcbaGaamyAaiaadshacqGHsislcaaIXaaabeaaaOqaamaabmaaba GaamyramaaBaaaleaacaWGPbGaamiDaaqabaGccqGHRaWkcaWGfbWa aSbaaSqaaiaadMgacaWG0bGaeyOeI0IaaGymaaqabaaakiaawIcaca GLPaaacaGGVaGaaGOmaaaacGaGRiilaaaa@4B1D@$

where ${E}_{it}$ and ${E}_{it-1}$ were the firm’s employment levels in year $t$ and $t-1$ , respectively.

This growth rate, described in the study by Davis, Haltiwanger and Schuh (1996), was bounded between -2 (exit) and 2 (entry). It was similar to the log growth rate for firms expanding or contracting employment by up to 50%. If the change in employment between consecutive years was greater than 0, the firm was a job creator. If the change in employment was less than 0, the firm was a job destroyer (i.e., shrinking employment). The employment change (either in level or growth rate) between each set of consecutive years for each firm constituted an observation. This was referred to as a firm-year observation. The characteristics of the firm and its owners were included with each firm-year observation.

The sample consisted of all firm-year observations belonging to privately owned Canadian-controlled corporations in the private sectorNote  that had employment at some point over the 2002/2003-to-2012/2013 period.

The first objective of this paper was to determine whether immigrant-owned private incorporated companies had lower or higher levels of job creation or destruction than firms with Canadian-born owners. The second objective of this paper was to determine whether immigrant-owned firms were more likely to be high-growth or rapidly shrinking firms than firms with Canadian-born owners.

Different samples were used to address each of these two objectives. To determine levels of job creation and destruction and net job change as part of the first objective, the sample consisted of all private incorporated companies with employment over the 2002/2003-to-2012/2013 period. This included entrants, incumbents (i.e., continuing firms) and exits, and produced a sample of 7.49 million firm-year observations (8.60 million with weights).

An entrant in year $t$ was a firm that existed during that year, but not the previous year. An exit in year $t$ was a firm that had employment in year   $t-1$ but not in year $t$ . For an entrant in year  $t$ , the gross job gain was the employment observed during that first year $t$ . This may have been a partial year since the firm may have started at any time during the year. Similarly, for an exit, the gross job loss was the employment observed during year $t-1$ . This, too, may have been a partial year, since the firm could have ceased operations at any time during the year.

To address employment growth rates as part of the second objective, the sample was restricted to incumbents. Entrants and exits were excluded because an entrant grows from zero employment to some finite value, and an exit falls from some finite value to zero employment. Calculating growth rates for such firms is difficult. The method used to calculate growth rates, following that of Davis, Haltiwanger and Schuh (1996), would produce growth rates of +2 for entrants and -2 for exits, if they were included.

This paper examined whether immigrant-owned firms or firms with Canadian-born owners were more likely to be rapidly growing firms, defined as firms with a growth rate greater than 0.2 (20%). All entrants would be seen as fast-growing firms, no matter how small their employment during the first year. Many entrants start with less than one full employee (ALU). Similarly, all exits would be seen as rapidly shrinking firms, no matter how small their employment during the last year.

For incumbent firms, a growth rate greater than 20% likely means a significant increase in employment. Therefore, the focus was on incumbents only when looking at the likelihood of being a rapidly growing or shrinking firm. Firms with an average size strictly smaller than one ALU over $t$ and $t-1$ were also excluded when the second objective was addressed, to avoid the small-size bias in growth rate calculations.Note  These smaller firms were included in the analysis of employment level changes, since a similar kind of bias does not apply. The sample size for the second objective was 4.60 million firm-year observations (5.27 million with weights).

To adjust the change in employment for differences in firm and owner characteristics between immigrant-owned firms and firms with Canadian-born owners, both ordinary least squares (OLS) and logit regressions were used.

OLS regression was used to examine the first objective. The dependent variable was the change in employment levels between years $t-1$ and $t$ . All firm-year observations were included, whether they experienced positive, negative or no change in employment. The independent variables used as controls included firm age measured in years (0, 1, 2, 3, 4 or 5, 6 or 7, 8 or 9, 10 to 14, 15 to 19, 20 or more), firm size measured in ALUs (less than 1, 1 to 4, 5 to 19, 20 to 49, 50 to 99, 100 to 499, 500 or more), two-digit industry of the North American Industry Classification System (NAICS) (18 levels, as shown in Table 1), the province where the firm employed most of its workers, a dummy variable to identify firms with multiple owners, the gender mix of owners (female only, male only, both), and the year of the observation. Also included was a dummy variable that was 1 if the firm was immigrant-owned, and 0 otherwise (Canadian-born owner). This variable’s coefficient was the point of interest, as it indicated the difference in the average annual absolute change in employment between immigrant-owned firms and firms with Canadian-born owners.

For the second objective—estimating the difference in likelihood of being a high-growth or rapidly shrinking firm—logistic regression was used to adjust for differences in firm characteristics. Two separate regressions were used. In the first, the dependent variable was a binary value equal to 1 if the firm was a high-growth firm and 0 otherwise. For the second, the dependent variable equalled 1 if the firm was a rapidly shrinking firm, and 0 otherwise. In both regressions, the independent variables were identical to those used in the OLS equation above. Based on the coefficients from the logistic regression, the marginal probabilities were estimated. As before, the focus was on the immigrant status variable, since it indicated the difference in the probability of an immigrant-owned firm or a firm with a Canadian-born owner being a high-growth firm or a rapidly shrinking firm.

As noted earlier, the same firm can contribute multiple observations since an observation is a firm-year. Unfortunately, it was not possible to determine how long a firm was in the sample, and therefore it was not possible to correct for the effect of this lack of independent observations on the standard errors. The standard errors were likely underestimated, since the number of independent observations was, in essence, overestimated. However, with many millions of observations, the sample size was so large that even if it were reduced by a factor of 10 after adjusting for clustered observations, the statistically significant effects would likely remain significant. To test this, the model was run 11 times and estimated on annual samples (e.g., using only the observations for 2002/2003, then for 2003/2004, and so on). When proceeding in this manner, there was no problem with the independence of observations. The annual results of the tests of significance produced essentially the same results as were observed when using the entire sample over all years.

## 3 Characteristics of the sample

Private incorporated firms tended to be small. Among active firms with employees (incumbents plus entrants), 71% had fewer than five employees (Table 1, last column). Only about 3% had more than 50 employees. Private incorporated firms tended to be clustered in five industries that accounted for about 50% of all private incorporated firms, including (starting from the largest) professional, scientific and technical services; construction; retail trade; accommodation and food services; and other services (except public administration).

There were some important differences between firms owned by immigrantsNote  and those owned by the comparison group, mainly Canadian-bornNote  owners. Immigrant-owned firms tended to be much younger. Among firms over the 2003-to-2013 period, more than three-quarters of immigrant-owned firms were younger than 10 years old (81%), compared with half of those with Canadian-born owners (53%). Twenty-three percent of firms with Canadian-born owners were over 20 years old, compared with 4% of immigrant-owned firms.

This study was particularly concerned with the share of young firms (4 years old or younger). During the period of study, the share of young firms was between 23 and 27 percentage points higher among immigrant-owned firms than among firms with Canadian-born owners. The difference averaged 25 percentage points (Table 1). The fact that immigrant-owned firms tend to be younger is important, since younger firms tend to grow faster than older firms (Haltiwanger, Jarmin and Miranda [2013], for the United States; Dixon and Rollin [2012 and 2014], for Canada).

Immigrant-owned firms also tended to be smaller than firms with Canadian-born owners. Eighty-one percent of immigrant-owned firms had fewer than five employees, compared with 69% of firms with Canadian-born owners. This result was consistent with other research (Fairlie and Lofstrom 2013; Green et al. 2016). Smaller firms tend to grow faster than larger firms, although—in absolute terms—larger firms create more jobs. Smaller firms are also more likely to shrink rapidly (Dixon and Rollin 2014).

The industrial distribution of immigrant-owned firms and firms with Canadian-born owners was similar, with a few exceptions. Immigrant-owned firms were more likely than firms with Canadian-born owners to be in accommodation and food services, and less likely to be in construction.

## 4 The number of jobs created and lost

Gross job creation and destruction far exceeded the net change in employmentNote  (see Davis, Haltiwanger and Schuh [1996] for the United States; and Rollin [2012] for Canada). For example, among the private incorporated firms studied over the 2003-to-2013 period, roughly 8.4 million jobs were created by entrants and expanding incumbents, and roughly 6.8 million jobs were lost by exits and contracting incumbents. This resulted in a net job change of 1.6 million jobsNote  or, more precisely, ALUs (Table 2-1 to 2-5). Note that ALU counts may differ from other employment estimates published by Statistics Canada (see “Methods” section).

There were 8.6 million weighted firm-year observations over the 11 years studied. Therefore, the annual average net number of jobs (ALUs) created per firm was 0.18 (1.6 million divided by 8.6 million). Among entering, exiting and incumbent firms with Canadian-born owners, there was a net change of about 1.2 million jobs. There were 7.2 million observations involving firms with Canadian-born owners, so the annual average net number of jobs created per firm was 0.16. Among immigrant-owned firms, roughly 0.4 million net jobs were created. There were 1.4 million observations involving immigrant-owned firms, so the annual average net number of jobs created per firm was 0.28 among immigrant-owned firms.

Therefore, 25% of the net new jobs created by private incorporated companies were attributable to immigrant-owned firms, although they accounted for only 17% of the private incorporated firms studied. Three-quarters of the net jobs created were attributable to firms with Canadian-bornNote  owners (since there were many more of them and they tended to be larger). Over the 11-year period, immigrant-owned firms tended to create more net new jobs on average (0.28) on a per-firm basis than firms with Canadian-born owners (0.16).

While accurate, results based on averages tell only part of the story. The majority of firms, whether immigrant-owned or with Canadian-born owners, generate few jobs or lose few jobs. Relatively few firms account for the majority of both gross job creation and gross job destruction. For example, among all firms (including entrants, incumbents and exits), the 10% of firms that were the largest job creators (i.e., with the highest annual employment change) accounted for 62% of gross job creation, with an average of 11.6 jobs created per firm. The top 5% accounted for 49% of gross jobs creation and created an annual average 18.3 jobs per firm (Table 3). The same was true for firms that lost jobs between consecutive years. The top 10% of job destroyers also accounted for 65% of all jobs lost, and the top 5% accounted for 52%. This concentration of job creation and destruction among relatively few firms was observed among both immigrant-owned firms and firms with Canadian-born owners. Of course, much of this pattern related to the size of the firms.

Finally, the gross job creation and destruction numbers quoted above tell a story of tremendous annual volatility in employment levels in firms. Net job creation is much smaller. As noted earlier, over the period studied, gross job creation registered 8.4 million ALUs and gross job loss 6.8 million ALUs, for a net job gain of 1.6 million ALUs. Similar patterns were observed for both immigrant-owned private incorporated companies and their counterparts with Canadian-born owners.

### 4.1 Adjusting for differences in firm characteristics

To adjust for differences in firm characteristics between immigrant-owned firms and firms with Canadian-born owners, the regression model described in Subsection 2.2 was applied. The dependent variable was the change in employment for each firm-year (i.e., the change in employment between $t$ and $t-1$ in each firm). All firm-year observations were in the sample, including those generated by entrants, exits and incumbent firms.

First, a model was run with only an immigrant status (immigrant or Canadian-born owner) dummy. The coefficient on this variable was 0.119 (statistically significant), which was the difference between immigrant firms and firms with Canadian-born owners in the average annual net employment generated per firm when no control variables were applied (i.e., the difference observed in the raw data, Tables 2-1 to 2-5). Then the control variables were added, which included firm age, firm size, industry, firm main province of employment, year of the observation, number of owners, gender mix of owners, and the main variable of interest—the immigrant status of owners. In this case, the coefficient on the immigrant status variable was -0.098 and was significant (Table 4).

The difference in the coefficient between the two regressions indicated the extent to which the control variables accounted for or “explained” the difference in the job creation behaviour of immigrant-owned firms and firms with Canadian-born owners. This difference was 0.217. A straightforward Oaxaca decomposition was used to decompose the difference in the net job creation per firm between immigrant-owned firms and firms with Canadian-born owners. As noted, 0.217 of this difference was “explained” by the control variables (more than 100% of the total). The decomposition demonstrated that 0.396 of the difference was because immigrant-owned firms were younger than their Canadian-born counterparts. Younger firms tend to be more dynamic and create and lose jobs at a higher rate than older firms. This represented more than all of the “explained” difference (0.271). The effect of firm age was partially offset by the effect of other independent variables.Note

To summarize, with regard to gross job change, immigrant-owned firms were more likely than firms with Canadian-born owners to be job creators than job destroyers (including incumbents, entrants and exits, Tables 2-1 to 2-5). As a result, in terms of net job change based on the unadjusted (raw) data, immigrant-owned firms held the advantage. On average, immigrant-owned firms accounted for 25% of the net jobs created over the 11-year period, while representing 17% of all private incorporated firms in the sample. This tendency to have higher net employment creation per firm was primarily because they were younger firms. When this difference was accounted for, the difference in average annual net job creation between immigrant-owned firms and firms with Canadian-born owners disappeared and, if anything, immigrant-owned firms had an annual average net job creation rate lower than that of firms with Canadian-born owners.

Why does that happen? In any given year, young firms (4 years old or younger) were much more likely to increase employment than to shed employment. Young firms accounted for 40.5% of gross job creation, but only 17% of gross job losses (Table 5) in the period studied. Even if entrants were excluded, since by definition they cannot shed employment, the tendency among young firms was very much toward gross job creation. Young incumbent firms accounted for 24% of gross job creation and 17% of gross job destruction. The opposite was true for older firms that had been in existence for over 20 years. Job losses among those firms outstripped job gains. Older firms accounted for 26% of all gross job gains, but 39% of gross job losses. In terms of net job gains, younger firms as a group accounted for the majority of the gains, and older firms as a group registered net job losses. While some older firms registered job gains, more registered job losses.

Immigration leads to the creation of new, young and dynamic private incorporated firms. This occurs because the companies owned by recent immigrants tend to be young (4 years old or younger) since immigrants have been in Canada for a shorter period of time than Canadian-born owners of the same age. In short, there is a “years since immigration” effect. This effect can be seen clearly in Chart 1. In 2012, 90% of firms owned by immigrants who had lived in Canada for five years were young firms. Even among firms owned by immigrants who had lived in Canada for 20 years, 43% were young. Only among immigrant owners who had lived in Canada for over 30 years did the proportion of young firms in 2012 approach that of Canadian-born owners, at roughly 29%.

Data table for Chart 1 ﻿
Data table for Chart 1
Table summary
This table displays the results of Data table for Chart 1. The information is grouped by Number of years since immigration (appearing as row headers), Immigrant-owned firms, calculated using percent units of measure (appearing as column headers).
Number of years since immigration Immigrant-owned firms
percent
5 90.2
6 83.0
7 78.3
8 73.2
9 69.2
10 66.2
11 63.0
12 60.2
13 58.1
14 56.9
15 52.8
16 50.9
17 48.0
18 47.3
19 45.9
20 43.5
21 42.4
22 41.7
23 40.7
24 38.8
25 37.2
26 34.7
27 33.8
28 33.3
29 31.5
30 32.5
31 29.6
32 30.7

But how much greater is the tendency among all immigrants in Canada to own young firms? Based on the sample of immigrants who have entered Canada since 1980, this study found that, over the entire 2003-to-2013 period, 58% of the private incorporated companies owned by immigrants were young firms (4 years old or younger), compared with about 32% among Canadian-born owners—a difference of 26 percentage points (Table 1). The share of young firms declined over the study period. From 2003 to 2012, this share fell for both immigrant-owned firms (from 60% to 54%) and firms with Canadian-born owners (from 35% to 29%). This finding was consistent with earlier research (MacDonald 2016; Decker et al. 2014; Criscuolo, Gal and Menon 2014). However, the difference between immigrant-owned firms and firms with Canadian-born owners remained at 25 percentage points. For the entire immigrant population, including those who have entered Canada prior to 1980, this may have been an overestimate of both the share of young immigrant-owned firms and the difference with firms with Canadian-born owners, since longer-term immigrants tend to own older firms. When all immigrant cohorts who have entered Canada since 1955 were included, it was estimated that 47% to 49% of all immigrant-owned firms in 2012 were young, well above the 29% observed among firms with Canadian-bornNote  owners—a difference of roughly 18 to 20 percentage points (Appendix).Note

## 5 The likelihood of being a high-growth firm

This section focuses on the employment growth rate of incumbent firms, and examines whether there is a difference between immigrant-owned continuing firms and continuing firms with Canadian-born owners in terms of likelihood of being a high-growth or rapidly shrinking firm. The most common definition of a high-growth firm comes from Eurostat and the Organisation for Economic Co-operation and Development (Eurostat and OECD 2007; Audretsch 2012). It defines high-growth firms as firms with 20% average annualized growth over three consecutive years.Note  Because this paper uses data on year-to-year changes, for any two consecutive years, high growth is defined as an annual growth rate of 20% or more. Rapidly shrinking firms are defined in this study as having an annual growth rate of -20% or less. Growth rates are estimated using the formula described in Susbection 2.2.

Entrants and exits were excluded since, in a comparison between immigrant-owned firms and firms with Canadian-born owners, restricting the sample to incumbents was more informative when it came to growth rate patterns, for reasons discussed in Subsection 2.2.

### 5.1 Growth rate distributions

The distribution of the growth rates showed that there were fewer immigrant-owned businesses than businesses with Canadian-born owners at the peak of the distribution (Chart 2). Therefore, there were more immigrant-owned firms present in the tails of the distribution, notably above 0.2 on the positive side. The difference was more pronounced on the job creation side of the distribution, and barely noticeable on the job loss side.

Data table for Chart 2 ﻿
Data table for Chart 2
Table summary
This table displays the results of Data table for Chart 2. The information is grouped by Growth rate (appearing as row headers), Canadian-born-owned and Immigrant-owned, calculated using frequency (percent) units of measure (appearing as column headers).
frequency (percent)
-1.95 0.0513 0.0327
-1.90 0.0566 0.0447
-1.85 0.0637 0.0576
-1.80 0.0666 0.0640
-1.75 0.0721 0.0704
-1.70 0.0772 0.0756
-1.65 0.0831 0.0838
-1.60 0.0867 0.0856
-1.55 0.0909 0.0860
-1.50 0.0992 0.1101
-1.45 0.1086 0.1126
-1.40 0.1155 0.1250
-1.35 0.1267 0.1368
-1.30 0.1380 0.1405
-1.25 0.1513 0.1567
-1.20 0.1685 0.1779
-1.15 0.1773 0.1868
-1.10 0.2020 0.2034
-1.05 0.2166 0.2395
-1.00 0.2441 0.2516
-0.95 0.2648 0.2864
-0.90 0.2970 0.3052
-0.85 0.3430 0.3596
-0.80 0.3830 0.3999
-0.75 0.4321 0.4594
-0.70 0.4930 0.5410
-0.65 0.5712 0.6258
-0.60 0.6510 0.6956
-0.55 0.7630 0.8101
-0.50 0.8918 0.9403
-0.45 1.0709 1.1064
-0.40 1.2932 1.3769
-0.35 1.6031 1.6816
-0.30 2.0067 2.0346
-0.25 2.5578 2.5277
-0.20 3.2793 3.1627
-0.15 4.3495 4.0675
-0.10 5.9119 5.3284
-0.05 8.0555 7.0085
0.00 10.8039 9.3609
0.05 10.3068 8.9284
0.10 8.4817 7.3705
0.15 6.5211 6.0424
0.20 4.9291 4.8878
0.25 3.7882 3.9226
0.30 2.9166 3.2498
0.35 2.2732 2.6537
0.40 1.8048 2.2031
0.45 1.4496 1.8230
0.50 1.1860 1.5172
0.55 0.9904 1.3029
0.60 0.8143 1.0961
0.65 0.7074 0.9691
0.70 0.6019 0.8408
0.75 0.5203 0.7306
0.80 0.4487 0.6510
0.85 0.3966 0.5764
0.90 0.3472 0.5229
0.95 0.3081 0.4824
1.00 0.2764 0.4413
1.05 0.2506 0.4051
1.10 0.2258 0.3494
1.15 0.2093 0.3361
1.20 0.1923 0.3158
1.25 0.1817 0.2883
1.30 0.1589 0.2611
1.35 0.1455 0.2512
1.40 0.1354 0.2357
1.45 0.1288 0.2282
1.50 0.1194 0.2100
1.55 0.1063 0.2029
1.60 0.0998 0.1808
1.65 0.0901 0.1715
1.70 0.0899 0.1608
1.75 0.0846 0.1528
1.80 0.0770 0.1360
1.85 0.0709 0.1272
1.90 0.0636 0.1230
1.95 0.0512 0.0855
1.99 0.0328 0.0470

Table 6 presents cumulative frequencies by growth rate. This table indicates the share of firms that are above or below any given growth rate. Among job creators (firms with positive employment growth), 27.3% of immigrant-owned firms were considered high-growth firms (i.e., growth rates above 0.2 or 20%) compared with 21.3% of firms with Canadian-born owners—a difference of 5.9 percentage points. Put another way, immigrant-owned firms were about 28%Note  more likely to be high-growth firms, an important difference. But the share of firms considered rapidly shrinking (growth rates below -0.2) was also marginally higher among immigrant-owned firms—19.8% versus 19.3%—a difference of 0.5 percentage points, or about 3% more likely to be rapidly shrinking. Immigrant-owned firms had a greater advantage over firms with Canadian-born owners on job creation (28% more likely) compared with job destruction (3% more likely). Table 6 shows that the differences on the positive growth side are both larger and present over a wider range of growth rates than on the negative growth side.

### 5.2 Adjusting for differences in firm characteristics

Some of these differences may be related to firm characteristics other than ownership, as noted earlier.

To adjust for these differences, logistic regression was used (see Subsection 2.2 for descriptions of these regressions). Two regression models were run: the first to estimate the likelihood of being a high-growth firm (sample of job creators), the second to estimate the likelihood of being a rapidly shrinking firm (sample of job destroyers). The independent (control) variables in both regressions included firm age, firm size, industry, province, number of owners, gender mix of owners and—most importantly—an immigrant status variable. The marginal probabilities of being a high-growth or rapidly shrinking firm were estimated based on the coefficients from the logistic regression.

The advantage that immigrant-owned firms held over firms with Canadian-born owners regarding the probability of being a high-growth firm decreased from 5.9 to 1.8 percentage points after adjusting for differences in firm characteristics (Table 7). The difference in the likelihood of being a rapidly shrinking firm was reduced from 0.5 to 0.4 percentage points. Immigrant-owned firms held an advantage in both, but it was considerably reduced, as expected.

To summarize, the unadjusted results (raw data) suggested that immigrant-owned firms were substantially more likely to be high-growth firms than firms with Canadian-born owners, and only marginally more likely to be rapidly shrinking. When the results were adjusted for differences in firm characteristics, the difference in the probability of being a high-growth or rapidly shrinking firm was reduced by 70%. Immigrant-owned firms were 1.8 percentage points, or 8%, more likely to be high-growth firms, and 0.4 percentage points, or 2%, more likely to be rapidly shrinking firms. These results for Canada were consistent with the results for the United States obtained by Kerr and Kerr (2016).

### 5.3 The effect of immigrant characteristics on being a high-growth firm

There is considerable heterogeneity among different types of immigrants regarding their tendency to be entrepreneurs and the types of businesses they own (Green et al. 2016; Fairlie and Lofstrom 2013). This study examined whether there was variation among incumbent firms regarding their tendency to be high-growth or rapidly shrinking firms, depending on the characteristics of the immigrant owners. These characteristics were defined in terms of immigrant class (economic, family, refugee, other and mixed [if there were multiple owners]), source region, educational attainment at arrival, and knowledge of an official language at arrival. Logistic regression was used. The independent variables included firm characteristics (firm age, firm size, industry and province), and the immigrant characteristics mentioned above.Note

#### Highest level of educational attainment among owners

Educational attainment of immigrant owners at the time of landing made little or no difference in the likelihood of an immigrant-owned firm being a high-growth or rapidly shrinking firm. The coefficients on the educational attainment variables in the logistic regressions were statistically or economically insignificant, which implies similar probabilities of being a high-growth or rapidly shrinking firm (Table 8).

#### Immigrant class

Similarly, the owner’s immigrant class appeared to make little difference regarding the probability of a firm being a high-growth or rapidly shrinking firm, with one exception. After the controls for firm and worker characteristics, firms owned by refugees had a marginally higher probability of being high-growth firms. The likelihood was roughly 1.0 percentage point higher among firms owned by refugees compared with the family and economic classes.

#### Source region

Firms owned by immigrants from China, India, North Africa and the Middle East, and Latin America had the highest probability of being high-growth firms. But the differences were not great (about 1.0-percentage-point difference with other source regions). There was also some small difference in the likelihood of being a rapidly shrinking firm based on the source region of the owner. Firms owned by immigrants from Western and Eastern Europe were less likely to shrink rapidly than firms owned by immigrants from Africa, North Africa and the Middle East, and Latin America (1.5-percentage-point difference).

Overall, the characteristics of immigrant owners (source region, educational attainment and immigrant class) had relatively little effect on the likelihood of a firm being a high-growth or rapidly shrinking firm.

## 6 Conclusion

This paper used microdata from the Canadian Employer $–$ Employee Dynamics Database to examine whether immigrant-owned private incorporated companies disproportionately contributed to job creation in Canada. Job creation was measured both by the number of jobs created and lost by firms, and by the employment growth rate.

The gross job creation and destruction analysis found that firms owned by immigrants who have entered Canada since 1980Note —the population of immigrants used in this study—were more likely than firms with Canadian-born owners to be job creators than job destroyers. This gave immigrant-owned firms an advantage in net job creation. As a result, the unadjusted (raw) data indicated that immigrant-owned private incorporated companies registered a higher average annual net job creation per firm over the decade than their counterparts with Canadian-born owners. Firms owned by immigrants accounted for a disproportionate share of net job creation; they accounted for 25% of net jobs created by private incorporated firms over the 11-year period, while representing 17% of the firms studied. The analysis also found that entering firms played a larger role in job creation among the immigrant-owned firm population than among the population of firms with Canadian-born owners. Also, the immigrant-owned firms in the sample accounted for a disproportionate share of entering firms when compared with firms with Canadian-born owners.

Immigrant-owned firms displayed significantly different characteristics than firms with Canadian-born owners. Most importantly, they tended to be younger, and younger firms create jobs at a higher rate than older firms. After adjusting for differences in firm age, firm size and other characteristics, the difference between immigrant-owned firms and firms with Canadian-born owners in the net job creation rate disappeared and, if anything, immigrant-owned firms had a lower annual net job creation rate.

The analysis also reaffirmed that, among immigrant-owned firms and firms with Canadian-born owners, both job creation and job loss were heavily concentrated among very few firms. The majority of private incorporated companies create or lose few jobs.

A similar trend emerged when focusing on a firm’s likelihood of being a high-growth firm (i.e., having over a 20% annual growth rate) or a rapidly shrinking firm (i.e., having a growth rate less than -20%). With regard to incumbent (i.e., continuing) firms, the unadjusted results indicated that immigrant-owned firms were 1.3 times more likely to be high-growth firms than firms with Canadian-born owners. There was little difference in the likelihood of being a rapidly shrinking firm. However, when the results were adjusted for differences in firm characteristics, the advantage held by immigrant-owned firms was reduced by 70%.

The characteristics of the immigrant owners (source region, educational attainment and immigrant class) had relatively little effect on the likelihood of a firm being a high-growth or rapidly shrinking firm.

## 7 Tables

﻿
﻿
Number of firm-year observations (weighted) Number of ALUs created Average job creation per firm-year Share of entry gross job creation 719,800 1.3 1.74 79.8 255,100 0.3 1.24 20.2 974,800 1.6 1.61 100.0 Note: ALU: average labour unit. Source: Statistics Canada, Canadian Employer–Employee Dynamics Database.
﻿
Number of firm-year observations (weighted) Number of ALUs created Average job creation per firm-year Share of incumbent gross job creation 2,962,600 6.0 2.02 88.3 564,100 0.8 1.41 11.7 3,526,700 6.8 1.92 100.0 Note: ALU: average labour unit. Source: Statistics Canada, Canadian Employer–Employee Dynamics Database.
﻿
Number of firm-year observations (weighted) Number of ALUs destroyed Average job destruction per firm-year Share of incumbent gross job destruction 2,907,800 5.2 1.79 90.3 487,700 0.6 1.15 9.7 3,395,500 5.8 1.70 100.0 Note: ALU: average labour unit. Source: Statistics Canada, Canadian Employer–Employee Dynamics Database.
﻿
Number of firm-year observations (weighted) Number of ALUs destroyed Average job destruction per firm-year Share of exit gross job destruction 550,300 0.9 1.58 85.4 129,000 0.1 1.15 14.6 679,300 1.0 1.49 100.0 Note: ALU: average labour unit. Source: Statistics Canada, Canadian Employer–Employee Dynamics Database.
﻿
Number of firm-year observations (weighted) Number of ALUs created Average net job growth per firm-year Share of net job growth 7,159,900 1.2 0.16 74.3 1,437,000 0.4 0.28 25.7 8,596,900 1.6 0.18 100.0 Note: ALU: average labour unit. Source: Statistics Canada, Canadian Employer–Employee Dynamics Database.
﻿
Gross job creation Gross job destruction Share of total Average job creation job creation per firm-year 100.0 2.0 100.0 1.8 100.0 1.4 100.0 1.1 100.0 1.9 100.0 1.7 77.2 7.6 79.0 6.9 72.6 4.9 74.9 4.3 76.7 7.1 78.6 6.5 63.0 12.4 65.5 11.5 57.3 7.8 60.1 6.9 62.4 11.6 65.1 10.8 50.0 19.7 52.8 18.6 43.8 11.9 47.0 10.8 49.4 18.3 52.4 17.4 Note: ALU: average labour unit. Source: Statistics Canada, Canadian Employer–Employee Dynamics Database.
﻿
Model without controls Model with controls 0.119Note *** -0.098Note *** Note ...: not applicable 0.116Note *** Note ...: not applicable -0.076Note *** Note ...: not applicable -0.112Note *** Note ...: not applicable 0.035 Note ...: not applicable 2.635Note *** Note ...: not applicable 1.919Note *** Note ...: not applicable 0.850Note *** Note ...: not applicable 0.741Note *** Note ...: not applicable 0.691Note *** Note ...: not applicable 0.623Note *** Note ...: not applicable 0.545Note *** Note ...: not applicable 0.470Note *** Note ...: not applicable 0.346Note *** Note ...: not applicable -3.677Note *** Note ...: not applicable -3.294Note *** Note ...: not applicable -2.860Note *** Note ...: not applicable -2.298Note *** Note ...: not applicable -1.650Note *** Note ...: not applicable -5.790 0.163Note *** 2.678Note *** 8,596,900 8,596,900 ... not applicable Note *** significantly different from reference category (p < 0.001) Return to note *** referrer Notes: These regressions include all firm-year observations (entrants, exits and incumbents). The model with controls also includes controls for the number of owners, the gender of the owners, the firm’s industry, the firm’s main province of employment and the year. For the model with controls, the reference category is Canadian-born-owned, single owner, male-only owners, firm age 20 years or older, firm size 100 to 499 average labour units (ALUs) or more, retail trade industry (North American Industry Classification System codes 44 and 45), main province Ontario and year 2013. Source: Statistics Canada, Canadian Employer–Employee Dynamics Database.
﻿
Share of gross job creation Share of gross job destruction Share of net job growth 14.6 Note ...: not applicable 90.5 22.5 15.3 59.9 14.9 17.7 0.2 11.3 14.5 -5.5 8.4 11.2 -6.1 28.3 41.3 -39.0 100.0 100.0 100.0 25.8 Note ...: not applicable 70.8 36.8 32.9 43.6 16.5 27.3 -2.3 8.7 15.4 -3.0 4.4 8.5 -2.8 7.7 15.8 -6.4 100.0 100.0 100.0 16.1 Note ...: not applicable 85.5 24.4 17.2 55.7 15.1 18.7 -0.4 10.9 14.6 -4.9 7.9 11.0 -5.2 25.6 38.6 -30.6 100.0 100.0 100.0 ... not applicable Notes: The numbers for gross job creation include firm-year observations with a strictly positive change in employment. The numbers for gross job destruction include firm-year observations with a strictly negative change in employment. The numbers for net job growth consider all firm-year observations, including those with unchanged employment. Percentages may not add up to 100% because of rounding. Source: Statistics Canada, Canadian Employer–Employee Dynamics Database.
﻿
Table 6
Cumulative frequency below and above selected growth rates, for incumbent (continuing) firms
Table summary
This table displays the results of Cumulative frequency below and above selected growth rates. The information is grouped by Growth rate (appearing as row headers), Cumulative frequency below growth rate and Cumulative frequency above growth rate, calculated using percent units of measure (appearing as column headers).
Growth rate Cumulative frequency below growth rate Cumulative frequency above growth rate
percent
-0.25 16.0 16.6 0.6 84.0 83.4 -0.6
-0.20Table 6 Note 1 19.3 19.8 0.5 80.7 80.2 -0.5
-0.15 23.6 23.8 0.2 76.4 76.2 -0.2
-0.10 29.6 29.2 -0.4 70.4 70.8 0.4
-0.05 37.6 36.2 -1.5 62.4 63.8 1.5
0.00 48.4 45.5 -2.9 51.6 54.5 2.9
0.05 58.7 54.4 -4.3 41.3 45.6 4.3
0.10 67.2 61.8 -5.4 32.8 38.2 5.4
0.15 73.7 67.9 -5.9 26.3 32.1 5.9
0.20Table 6 Note 2 78.7 72.7 -5.9 21.3 27.3 5.9
0.25 82.4 76.7 -5.8 17.6 23.3 5.8
0.75 95.7 93.1 -2.7 4.3 6.9 2.7
1.00 97.5 95.7 -1.8 2.5 4.3 1.8
1.25 98.5 97.4 -1.1 1.5 2.6 1.1
﻿
Probability of being a high-growth firm Probability of being a rapidly shrinking firm 21.9 19.3 23.7Note *** 19.7Note *** 36.5Note *** 17.6Note *** 21.2Note *** 20.2Note *** 18.0Note *** 20.2Note *** 16.1Note *** 20.0Note *** 13.3 19.7 22.9Note *** 23.1Note *** 21.6Note ** 15.8Note *** 19.2 11.7Note *** 18.2 10.0 18.4 9.5 19.0 9.3 5,273,100 5,273,100 Note ** significantly different from reference category (p < 0.01) Return to note ** referrer Note *** significantly different from reference category (p < 0.001) Return to note *** referrer Notes: Incumbent (continuing) firms with an average size of 1 average labour unit (ALU) or more in years t-1 and t only. High growth is defined as having an employment growth rate of 0.2 or more. Rapidly shrinking is defined as having an employment growth rate of -0.2 or less. The probabilities presented are average marginal probabilities over the sample, based on logistic regression. The probabilities were multiplied by 100. The logistic regression also included controls for the number of owners, the gender of the owners, the firm's industry, the firm’s main province of employment and the year. Source: Statistics Canada, Canadian Employer–Employee Dynamics Database.
﻿

## Appendix Estimating the proportion of immigrant-owned private incorporated companies aged 4 or younger in 2012

Because of data limitations, the analysis in the main body of this report was based on the population of immigrants who have entered Canada since 1980. As demonstrated earlier, the probability of immigrant-owned private incorporated companies being young firms (i.e., 4 years of age or younger) declines rapidly with the number of years the owner has been in Canada (Chart 1). By restricting the population to immigrants who have entered Canada since 1980, the proportion of young firms in the entire immigrant population was overestimated (i.e., if immigrants entering Canada after 1980 were included). The belief that the share of young private incorporated companies was greater among immigrant-owned firms than among firms with Canadian-born owners was essential to the conclusion that immigrant-owned firms contributed disproportionately to net job creation. Therefore, it was necessary to determine whether this held true for the entire population of immigrant-owned firms, not just those owned by immigrants who have entered Canada since 1980. This was determined by calculating the share of young firms in 2012, when the estimate was based on all immigrants who have entered Canada since 1955. Very few immigrants who entered Canada prior to 1955 owned firms in 2012.Note

To achieve this, five-year immigrant entry cohorts were defined from the 1955-to-1959 cohort to the 2005-to-2009 cohort, and the total number of immigrants entering Canada in each five-year cohort was determined (IRCC n.d.). The age distribution upon entry to Canada of each of these cohorts was then determined, and the cohort was “aged” to determine each cohort’s age distribution in 2012. Based on historical data and the relative probability of immigrants from different age groups (i.e., 25 to 34, 35 to 44, 45 to 54, 55 to 64) owning a firm, the likelihood of immigrants aged 25 to 64 owning a firm in 2012 was estimated for each of the cohorts from the 1955-to-1959 cohort to the 1975-to-1979 cohort. This determined the total number of immigrant-owned firms in 2012 from each entering cohort. Again based on historical data, the share of these young firms was estimated by applying two different scenarios. These steps determined the total number of private incorporated companies and the number of young firms owned in 2012 by immigrants from all entering cohorts since 1955. The details follow.

This study used the age distribution of entering immigrants, which was based on official statistics produced annually by Immigration, Refugees and Citizenship Canada (previously Citizenship and Immigration Canada, see Table A.1). For the cohorts in the 2000s, the 2005 age distribution (IRCC n.d.) was applied. For the cohorts in the 1990s, the 1994 age distribution—the oldest age distribution that could be located—was used (CIC 2003). The 1994 age distribution of entering immigrants was also applied to all the cohorts from the 1955-to-1959 cohort to the 1985-to-1989 cohort.

﻿
0 to 4 5 to 14 15 to 24 years 25 to 34 years 35 to 44 years 45 to 54 years 55 to 64 years 65 years or older years years 5.2 15.5 17.8 27.7 14.0 10.8 4.2 4.6 7.0 15.0 15.4 31.5 18.4 7.9 3.1 1.7 Sources: Immigration, Refugees and Citizenship Canada, n.d., Facts and Figures 2015: Immigration Overview - Permanent Residents, and Citizenship and Immigration Canada, 2003, Facts and Figures 2002: Immigration Overview.

Each cohort was then “aged” to determine the age distribution of that cohort in 2012 and the number of immigrants from each cohort who were aged 25 to 64 in 2012.

To determine the number of private incorporated companies in 2012 that were owned by immigrants from each cohort, the following approach was used.

The objective is to estimate $S$—the share of immigrant-owned private incorporated companies that were 4 years old or younger in 2012—for two different populations: all immigrants who have entered Canada since 1980, and all immigrants who have entered Canada since 1955.

For each entry cohort $i$ :

${N}_{i}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\left({R}_{i}\text{\hspace{0.17em}}\cdot \text{\hspace{0.17em}}{P}_{i}\right)$

${Y}_{i}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\left({N}_{i}\text{\hspace{0.17em}}\cdot \text{\hspace{0.17em}}{K}_{i}\right)$

where

${N}_{i}$ : the number of firms owned in 2012 by immigrants aged 25 to 64 from cohort $i$

${R}_{i}$ : the number of immigrants from entry cohort $i$ aged 25 to 64 in 2012

${P}_{i}$ : the proportion of the immigrant population aged 25 to 64 that owned a firm in 2012 (projected values for cohorts from the 1955-to-1959 cohort to the 1975-to-1979 cohort, see below)

${Y}_{i}$ : the number of firms in entry cohort $i$ that are 4 years old or younger in 2012

${K}_{i}$ : the proportion of all firms from entry cohort $i$ that are 4 years old or younger in 2012 (projected values for cohorts from the 1955-to-1959 cohort to the 1975-to-1979 cohort, see below).

### Projecting ${P}_{i}$ and ${K}_{i}$ for the entering cohorts from the 1955-to-1959 cohort to the 1975-to-1979 cohort

Rather than arbitrarily projecting ${P}_{i}$ (the proportion of the immigrant population aged 25 to 64 from cohort $i$ that owned a firm in 2012), available information on the characteristics of immigrant owners of private incorporated companies was used to inform the projection.

A paper by Green et al. (2016) documented the prevalence of ownership by immigrants of private incorporated companies in 2010. They found that age was a major determinant: 45- to 54-year-olds were twice as likely as 25- to 34-year-olds to own a company. For example, in 2012, immigrants from the 1960s cohorts are going to be older than those from the 1990s cohorts. A method of accounting for this age difference was developed in the present projection. Green et al. (2016) also concluded that gender was a factor, but since the distribution by gender of entering immigrants changed little from the 1960s to the 2000s, this variable would have little effect on the results in this paper.

The projection was adjusted to account for the fact that the age distribution in 2012 of immigrants from the earlier cohorts would be very different from that of the later cohorts. The relative (to immigrants aged 15 to 24) probabilities of immigrants owning a firm were as follows: ages 15 to 24, 1.0; ages 35 to 44, 1.9 times as likely; ages 45 to 54, 2.3 times as likely; ages 55 to 64, 1.6 times as likely (Green et al. 2016).

For any particular cohort $i$ , the relative probability of all immigrants aged 25 to 64 owning a firm in 2012 was the weighted average of the relative probabilities by age group, where the weights were the share of the population in each age group in 2012. That is,

$R P i = Σ k = 1 4 ( P k ⋅ r p k ) , MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaamOuaiaadc fadaWgaaWcbaGaamyAaaqabaGccaaMc8Uaeyypa0JaaGPaVpaaxada baGaeu4OdmfaleaacaWGRbGaeyypa0JaaGymaaqaaiaaisdaaaGcca aMc8UaaiikaiaaccfadaWgaaWcbaGaam4AaaqabaGccaaMc8UaeyyX ICTaaGPaVlaadkhacaWGWbWaaSbaaSqaaiacob4GRbaabeaakiaacM cacaGGSaaaaa@510E@$

where

$R{P}_{i}$ : relative probability of an immigrant from cohort $i$ owning a firm in 2012

$r{p}_{k}$ : relative probability of an immigrant in age group $k$ owning a firm

${P}_{k}$ : proportion of immigrants aged 25 to 64 in 2012 who are in age group $k$ .

The weights (the age distributions in 2012) are shown for all cohorts in Table A.2 below.

﻿
Table A.2
Estimating the relative probability of owning a firm in 2012, by immigrant cohort
Table summary
This table displays the results of Estimating the relative probability of owning a firm in 2012. The information is grouped by Entry cohort (appearing as row headers), Age distribution of 25- to 64-year-olds in 2012, Relative probability of a 25- to 64-year-old owning a firm in 2012, Proportion of 25- to 64-year-olds owning a firm in 2012, 25 to 34, 35 to 44, 45 to 54 and 55 to 64, calculated using percent, ratio and proportion units of measure (appearing as column headers).
Entry cohort Age distribution of 25- to 64-year-olds in 2012 RelativeTable A.2 Note 1 probability of a 25- to 64-year-old owning a firm in 2012 Proportion of 25- to 64-year-olds owning a firm in 2012
25 to 34 35 to 44 45 to 54 55 to 64
percent ratio proportion
2005 to 2009 36.1 32.7 19.4 11.9 1.62 0.026
2000 to 2004 25.4 39.6 19.4 15.4 1.70 0.037
1995 to 1999 22.7 31.5 28.6 16.9 1.75 0.033
1990 to 1994 20.7 23.8 37.1 18.2 1.31 0.028
1985 to 1989 17.6 22.7 31.3 28.2 1.77 0.034
1980 to 1984 7.8 23.5 26.8 41.0 1.80 0.040
1975 to 1979 0.0 24.6 31.7 43.3 1.89Table A.2 Note  0.042Table A.2 Note
1970 to 1974 0.0 13.5 40.3 46.2 1.92Table A.2 Note  0.043Table A.2 Note
1965 to 1969 0.0 0.0 43.7 56.2 1.90Table A.2 Note  0.043Table A.2 Note
1960 to 1964 0.0 0.0 25.1 74.9 1.77Table A.2 Note  0.039Table A.2 Note
1955 to 1959 0.0 0.0 0.0 100.0 1.60Table A.2 Note  0.036Table A.2 Note

The projected variable ${P}_{i}$—the proportion of 25- to 64-year-olds who owned a firm in 2012—was calculated such that it was proportional to the projected values of the relative probability of owning a firm in 2012 (i.e., $R{P}_{i}$ ) (se last column of Table A.2).

The second variable to be projected was ${K}_{i}$ , the proportion of all firms that were 4 years of age or younger. Two alternative projections were produced since there was little guidance. The first projection (projection A) assumed that the tendency of ${K}_{i}$ was to continue to decline with years since immigration. Therefore, the values declined from 0.250 to 0.120 between the 1975-to-1979 entering cohort and the 1955-to-1959 entering cohort (Table A.3). The second projection (projection B) assumed that ${K}_{i}$ stabilized in value 28 to 33 years after immigration, as demonstrated in Chart 1, and therefore ${K}_{i}$ remained at 0.313 for all cohorts prior to the 1980s. The outcome variable $S$—the share of immigrant-owned firms that were young in 2012—was minimally affected by these two different assumptions, varying by only 1 percentage point.

### The outcome

Once the two key ratios used in this estimation procedure were projected, the calculations to determine the primary outcome variable—the proportion of immigrant-owned firms that were young (4 years old or younger) in 2012—are shown below in Table A.

Based on immigrants who have entered Canada since 1980, 55% of immigrant-owned firms were 4 years old or younger in 2012. Based on the population of immigrants who have entered Canada since 1955, this proportion was 47% or 49%, depending on which projection of ${K}_{i}$ is used.

Approximately 29% of private incorporated companies owned by the Canadian-born population in 2012 were 4 years old or younger. Therefore, private incorporated companies owned by all immigrants were more likely than firms with Canadian-born owners to be young in 2012.

﻿
Table A.3
Estimating the proportion of immigrant-owned firms aged 4 or younger in 2012
Table summary
This table displays the results of Estimating the proportion of immigrant-owned firms aged 4 or younger in 2012. The information is grouped by Cohort i (appearing as row headers), Number of entering immigrants, Number of entering immigrants aged
25 to 64 in 2012 (Ri), Age range in 2012, Proportion of Ri owning a firm in 2012 (Pi), Number of firms owned in 2012 (Ni=Pi · Ri), Proportion of all firms aged 4 or younger in 2012 (Ki), Number of firms aged 4 or younger
(Yi=Ni · Ki), probability, projection, projection A and projection B, calculated using thousands, years and proportion units of measure (appearing as column headers).
Cohort i Number of entering immigrants Number of entering
immigrants aged
25 to 64 in 2012
(Ri)
Age range in 2012 Proportion of Ri owning a firmTable A.3 Note 1 in 2012
(Pi)
Number of firms
owned in 2012
(Ni=Pi · Ri)
Proportion of all firms
aged 4 or younger in 2012
(Ki)
Number of firms
aged 4 or younger
(Yi=Ni · Ki)
probability projection probability projection probability projection ATable A.3 Note 2 projection BTable A.3 Note 3 probability projection ATable A.3 Note 2 projection BTable A.3 Note 3
thousands years proportion thousands proportion thousands
2005 to 2009 1,249.1 904.9 25 to 64 0.026 Note ...: not applicable 23.6 Note ...: not applicable 0.864 Note ...: not applicable Note ...: not applicable 20.4 Note ...: not applicable Note ...: not applicable
2000 to 2004 1,164.1 812.5 25 to 64 0.037 Note ...: not applicable 30.4 Note ...: not applicable 0.659 Note ...: not applicable Note ...: not applicable 20.0 Note ...: not applicable Note ...: not applicable
1995 to 1999 1,018.9 797.7 25 to 64 0.033 Note ...: not applicable 26.5 Note ...: not applicable 0.533 Note ...: not applicable Note ...: not applicable 14.1 Note ...: not applicable Note ...: not applicable
1990 to 1994 1,184.8 1,016.5 25 to 64 0.028 Note ...: not applicable 28.3 Note ...: not applicable 0.440 Note ...: not applicable Note ...: not applicable 12.5 Note ...: not applicable Note ...: not applicable
1985 to 1989 688.7 506.9 25 to 64 0.034 Note ...: not applicable 17.4 Note ...: not applicable 0.376 Note ...: not applicable Note ...: not applicable 6.5 Note ...: not applicable Note ...: not applicable
1980 to 1984 570.2 380.9 30 to 64 0.040 Note ...: not applicable 15.4 Note ...: not applicable 0.313 Note ...: not applicable Note ...: not applicable 4.8 Note ...: not applicable Note ...: not applicable
1975 to 1979 650.4 344.1 35 to 64 Note ...: not applicable 0.042 Note ...: not applicable 14.4 Note ...: not applicable 0.250 0.313 Note ...: not applicable 3.6 4.5
1970 to 1974 799.3 301.3 40 to 64 Note ...: not applicable 0.043 Note ...: not applicable 12.9 Note ...: not applicable 0.220 0.313 Note ...: not applicable 2.8 4.0
1965 to 1969 912.2 271.8 45 to 64 Note ...: not applicable 0.043 Note ...: not applicable 11.6 Note ...: not applicable 0.180 0.313 Note ...: not applicable 2.1 3.6
1960 to 1964 456.2 95.3 50 to 64 Note ...: not applicable 0.039 Note ...: not applicable 3.7 Note ...: not applicable 0.150 0.313 Note ...: not applicable 0.5 1.2
1955 to 1959 788.5 104.1 55 to 64 Note ...: not applicable 0.036 Note ...: not applicable 3.7 Note ...: not applicable 0.120 0.313 Note ...: not applicable 0.4 1.2

## References

Anderson, S. 2016. Immigrants and Billion Dollar Startups. March. National Foundation for American Policy. NFAP Policy Brief. (accessed December 21, 2017).

Audretsch, D.B. 2012. Determinants of High-growth Entrepreneurship. Report prepared for the Organisation for Economic Co-operation and Development / Danish Business Authority international workshop “High-growth firms: Local policies and local determinants,” Copenhagen, March 28, 2012.

Audretsch, D.B., L. Klomp, E. Santarelli, and A.R. Thurik. 2004. “Gibrat’s Law: Are the services different?” Review of Industrial Organization 24 (3): 301–324.

CIC (Citizenship and Immigration Canada). 2003. Facts and Figures 2002: Immigration Overview. (accessed January 29, 2018).

Criscuolo, C., P.N. Gal, and C. Menon. 2014. The Dynamics of Employment Growth: New Evidence from 18 Countries. OECD Science, Technology and Industry Policy Papers, no. 14. Paris: OECD Publishing.

Davis, S.J., J. Haltiwanger, and S. Schuh. 1996. Job Creation and Destruction. Cambridge, Massachusetts: MIT Press.

Decker, R., J. Haltiwanger, R.S. Jarmin, and J. Miranda. 2014. The Secular Decline in Business Dynamism in the U.S. University of Maryland. Manuscript.

Decker, R.A., J. Haltiwanger, R.S. Jarmin, and J. Miranda. 2015. Where Has All the Skewness Gone? The Decline in High-growth (Young) Firms in the U.S. NBER Working Paper Series, no. 21776. Cambridge, Massachusetts: National Bureau of Economic Research.

Dixon, J., and A.-M. Rollin. 2012. Firm Dynamics: Employment Growth Rates of Small Versus Large Firms in Canada. The Canadian Economy in Transition, no. 25. Statistics Canada Catalogue no. 11-622-M. Ottawa: Statistics Canada.

Dixon, J., and A.-M. Rollin. 2014. The Distribution of Employment Growth Rates in Canada: The Role of High-growth and Rapidly Shrinking Firms. Economic Analysis Research Paper Series, no. 91. Statistics Canada Catalogue no. 11F0027F. Ottawa: Statistics Canada.

Dixon, J., R. Petrunia and A.-M. Rollin. 2018. “Studying Firm Growth Distributions with a Large Administrative Employment Database.” Journal of Economics and Statistics 238 (3-4): 189–221.

Eurostat and OECD (Organisation for Economic Co-operation and Development). 2007. Eurostat-OECD Manual on Business Demography Statistics. (accessed December 21, 2017).

Fairlie, R.W., and Lofstrom, M. 2013. Immigration and Entrepreneurship. IZA Discussion Papers, no. 7669. Bonn, Germany: Institute for the Study of Labor.

Geroski, P.A. 2005. “Understanding the implications of empirical work on corporate growth rates.” Managerial and Decision Economics 26 (2): 129–138.

Green, D., H. Liu, Y. Ostrovsky, and G. Picot. 2016. Immigration, Business Ownership and Employment in Canada. Analytical Studies Branch Research Paper Series, no. 375. Statistics Canada Catalogue no. 11F0019M. Ottawa: Statistics Canada.

Haltiwanger, J., R. Jarmin, and J. Miranda. 2013. “Who creates jobs? Small vs. large vs. young.” Review of Economics and Statistics 95: 347–361.

IRCC (Immigration, Refugees and Citizenship Canada). n.d. Facts & Figures 2015: Immigration Overview - Permanent Residents. (accessed January 29, 2018).

Kerr, S.P., and W.R Kerr. 2016. Economic Impacts of Immigration: A Survey. NBER Working Paper Series, no. 16736. Cambridge, Massachusetts: National Bureau of Economic Research.

Kerr, S.P., and W.R. Kerr. 2016. Immigrant Entrepreneurship. NBER Working Paper Series, no. 22385. Cambridge, Massachusetts: National Bureau of Economic Research.

Lafrance, A., and D. Leung. 2009. T2-LEAP (1984-2007): A Longitudinal Database of Incorporated Firms in Canada. Ottawa: Statistics Canada. Unpublished.

MacDonald, R. 2016. Business Entry and Exit Rates: A 30-year Perspective. Economic Insights, no. 38. Statistics Canada Catalogue no. 11-626-X. Ottawa: Statistics Canada.

OECD (Organisation for Economic Co-operation and Development). 2010. Open for Business: Migrant Enterprises in OECD Countries. Paris: OECD Publishing.

Ostrovsky, Y., G. Picot, and D. Leung. 2018. “The financing of immigrant-owned firms in Canada.” Small Business Economics.

Rollin, A.-M. 2012. Firm Dynamics: Employment Dynamics Arising from Firm Growth and Contraction in Canada, 2001 to 2009. The Canadian Economy in Transition, no. 24. Statistics Canada Catalogue no. 11-622-M. Ottawa: Statistics Canada.

Schuetze, H.J., and H. Antecol. 2007. “Immigration, entrepreneurship and the venture start-up process.” In The Life Cycle of Entrepreneurial Ventures, ed. S. Parker, International Handbook Series on Entrepreneurship, Vol. 3, p. 107–135. Boston: Springer Science and Business Media, Inc.

Is something not working? Is there information outdated? Can't find what you're looking for?

Privacy notice

Date modified: