How Canadian firms work: evidence from the 2024 Enterprise Survey
What holds Canadian businesses back, who innovates and exports, how firms finance themselves, and how this differs by size, sector, region and gender.
By Shadi Rahimfallah
This is an independent analysis. All analysis and interpretation are my own and are not affiliated with or endorsed by any company or organization, including the World Bank or any firm in the survey. Any errors are mine.
Data: World Bank Enterprise Survey, Canada 2024Formal private firms with 5+ employeesFieldwork April 2024 – April 2025
4parts
21sections
35further analyses
47figures
Part 1
Overview
What the survey is and the main messages.
1.1Key takeaways
Skills are the top constraint. An inadequately educated workforce is named as the biggest obstacle more often than anything else, ahead of tax rates and political instability.
Finance is not the bottleneck for most firms. Most firms that did not borrow say they did not need to; credit constraints are concentrated among small firms.
Canada trades with one partner. The large majority of exporters sell mainly to the United States, which is also the main source of imported inputs.
R&D is the common thread. Firms that spend on R&D are clearly more likely both to export and to innovate, after accounting for sector, region and size.
Small firms carry the heaviest compliance load. Per worker, tax compliance takes far more time in small firms than in large ones.
Regional gaps are small once industry mix is held constant. After sector and size controls, the provinces look very similar.
1.2What the survey covers
The Enterprise Survey interviews owners and managers of formal private firms about their day-to-day experience. For applied work, each question is a statistic about the business economy: a share of firms, a distribution, an average percentage, or a typical amount of time or money. Grouped by subject, the questions cover six areas of the economy.
Figure 1Questions by area of the economyNumber of questions with economic content in each area (survey identifiers and weights excluded).
Several topics cut across these areas: gender (ownership, management, workforce, training), the state (government ownership, contracts, inspections, state banks), formality and informality (informal competitors, informal payments), resources and utilities (energy, water, land, materials) and foreign links (ownership, licensed technology, exports, imported inputs). Results can be broken down by firm size, eight sectors, five regions and four city sizes.
Part 2
Descriptive findings
Weighted shares of Canadian firms by size, sector and region.
2.1What holds firms back
Firms picked the single biggest obstacle to their operations from fifteen options, and separately rated each one from “no obstacle” to “very severe”.
Figure 2The biggest obstacleThe eight most-named obstacles; the top three are highlighted.
Figure 3Major obstacles by firm sizeEach obstacle rated separately; dots show the share of firms rating it major or very severe.
The skills gap is felt across all sizes. Access to finance and the cost of hiring weigh more on small firms, while large firms point more often to tax rates.
2.2Innovation and R&D
Innovation means introducing a new or significantly improved product or process in the last three years. R&D covers in-house or contracted research, excluding market research.
Figure 4Innovation and R&D by firm sizeShare of firms introducing new products or processes, and spending on R&D.
Innovation is spread across firm sizes, but R&D spending rises sharply with size.
2.3Trade
Figure 5Exporters by sectorFirms exporting at least 10% of sales; the three most export-oriented sectors are highlighted.
Figure 6Main trade partnersMain export destination named by exporters and main source named by direct importers.
Manufacturing sectors and professional services are the most export-oriented. On both sides of trade the United States dominates, which makes Canadian firms highly exposed to U.S. trade policy.
2.4Finance
Figure 7Access to credit by firm sizeCredit constrained: rejected, or did not apply for reasons other than having enough capital.
Most firms finance working capital internally. Credit constraints are rare overall but more common among small firms, which also rate access to finance as a bigger obstacle.
2.5Gender
Figure 8Women in ownership and managementShare of firms with at least one female owner, and with a female top manager.
Women’s participation is highest in food and retail and lowest in fabricated metal. The gap between ownership and management is widest in food and professional services.
2.6Regions and infrastructure
Figure 9How the four main regions compareDarker cells mean a higher share than the other regions on that row; the territories are left out because their sample is very small.
Figure 10Infrastructure disruptions by city sizeShare of firms with at least one power outage or internet disruption in the last fiscal year.
Central Canada leads on R&D, credit, training and emissions monitoring; British Columbia on exporting; Atlantic Canada on innovation but also on outages and skills shortages.
2.7Green economy
Figure 11Green practices by firm sizeEnergy-management measures and CO₂ monitoring over the last three years.
Green practices rise steadily with firm size. Energy management is far more common than CO₂ monitoring, which remains a minority practice even among large firms.
2.8What goes together
Exporting, innovating and productivity are regressed on firm characteristics with sector and region controls, survey weights and robust standard errors. Blue dots have 95% intervals that exclude zero.
Figure 12Firm traits associated with each outcomeExporting and innovating: linear probability models (percentage points). Sales per worker: OLS on the log, shown as % difference.
Exporters are more often foreign-owned, certified and R&D-active. Innovation goes with R&D and with a credit line. Older, multi-site firms and firms with a website report higher sales per worker. Women-led firms report lower sales per worker within the same sector and region, an association that may reflect differences in activity within sectors.
The survey is a single cross-section, so these are associations, not causal effects.
Part 3
Further analyses
Thirty-five applied-economics analyses, grouped by topic. Each states the question, the finding, a figure and the method.
3.1Innovation
What drives innovation and R&D, and whether they pay off.
A01Regression
What makes a Canadian firm innovate?
Which firm traits go with introducing new products or processes?
R&D and credit go with innovation.
Spending on R&D and having a bank loan or credit line are the clearest correlates of both new products and new processes. Women-led and larger firms are slightly less likely to report new products, while foreign ownership, exporting, certification and training show no clear link once these are held constant. The two kinds of innovation are only weakly related to each other.
Figure 13What makes a Canadian firm innovate?Chart: average marginal effects with 95% intervals; faded marks are not statistically significant.
Method and data
Method
Probit or logit with survey weights, sector and region fixed effects; report average marginal effects. Run product and process innovation separately, then a bivariate probit to let the two decisions be correlated.
Questions used
h1h5l1b5b2bd3ch8l10k82b7b7ab8c22be6
Watch out
Cross-section: innovation and R&D are chosen together, so read effects as associations.
A02Advanced
Innovation and productivity: the CDM chain
Does R&D lead to innovation, and does innovation raise productivity?
The R&D → innovation → productivity chain is weak in Canada’s data.
Larger firms and exporters are much more likely to do R&D. But in the next steps, predicted R&D does not significantly raise the chance of innovating, and predicted innovation does not significantly raise sales per worker. Among R&D performers, foreign-owned firms spend more relative to sales.
Figure 14Innovation and productivity: the CDM chainChart: step 1 of the three-step CDM model (who does R&D). Steps 2 and 3 are reported in the text.
Method and data
Method
Crépon–Duguet–Mairesse model in three steps: (1) Heckman selection for whether and how much a firm spends on R&D (h8, h9/sales); (2) probit of innovation on predicted R&D intensity; (3) OLS of log sales per worker on predicted innovation.
Questions used
h8h9h1h5d2l1b2bd3ck82e2bl10
Watch out
Needs exclusion restrictions in step 1; R&D amounts (h9) are answered by fewer firms.
A03Regression
Does competition push or block innovation?
Is there an inverted-U between competition and innovation?
No inverted U between competition and innovation.
Neither the number of competitors nor its square is statistically significant, so the data do not support the inverted-U pattern found elsewhere. Firms that face informal competitors are somewhat more likely to innovate, but only at the margin of significance.
Figure 15Does competition push or block innovation?Chart: predicted chance of innovating across the number of competitors; the shape is not statistically significant.
Method and data
Method
Logit of innovation on number of competitors and its square (Aghion et al. 2005), plus pricing power (e33) and informal competition (e11). Plot predicted probability against competitors.
Questions used
h1h5e2be32e33e11e1
Watch out
Competitor counts are self-reported and top-coded answers are common; consider log or categories.
A04Regression
Foreign technology and in-house R&D: substitutes or complements?
Do firms that license foreign technology do more or less of their own R&D?
Foreign technology neither replaces nor clearly complements own R&D.
Licensing foreign technology has no significant link with doing R&D, for domestic or foreign-owned firms. Exporting is the strong correlate of R&D. The joint model finds no significant correlation between licensing and R&D decisions.
Figure 16Foreign technology and in-house R&D: substitutes or complements?Chart: average marginal effects on the probability of doing R&D.
Method and data
Method
Probit of R&D on foreign licence and interaction with foreign ownership; or bivariate probit of R&D and licensing.
Questions used
h8e6b2bd12bd3cl1
Watch out
Small number of licensing firms in some sectors.
3.2Productivity & growth
Who is productive, who grows, and why.
A05Regression
Exporter productivity premium
Are exporters more productive, and where in the distribution is the gap largest?
The exporter premium sits at the top of the distribution.
On average, exporters are not significantly more productive once foreign ownership, certification and R&D are controlled, and matching estimates are close to zero. Quantile regressions show why: the premium is not significant in the lower part of the distribution, then becomes significant from the median and is largest among the most productive firms.
Figure 17Exporter productivity premiumChart: exporter premium in sales per worker at each percentile, with 95% band (quantile regressions, which are unweighted).
Method and data
Method
OLS and quantile regressions (10th, 50th, 90th percentiles) of log sales per worker on an exporter dummy; propensity-score matching as a robustness check.
Questions used
d2l1d3bd3cb2bb5b8h8a4aa2a6a
Watch out
Self-selection: productive firms start exporting, so this is not the effect of exporting itself.
A06Advanced
Total factor productivity in manufacturing
How productive are manufacturing firms once capital and materials are accounted for?
Manufacturing output is driven by labour; TFP has no clear correlates.
In a Cobb-Douglas production function, the labour elasticity dominates, materials matter less and capital is close to zero, with roughly constant returns to scale. Productivity (TFP) is not linked to foreign ownership, exporting, R&D or age in this small manufacturing sample; training shows only a borderline negative association.
Figure 18Total factor productivity in manufacturingChart: output elasticities (left) and correlates of TFP (right).
Method and data
Method
Estimate a Cobb-Douglas (log sales on log labour cost, log materials, log capital replacement value) by sub-sector; keep the residual as TFP; then relate TFP to ownership, management and exporting.
Questions used
d2n2an2en7al1f1
Watch out
Manufacturing sub-sample only; capital is replacement value, not book value.
A07Regression
Does firm growth depend on size? (Gibrat’s law)
Do small firms grow faster than large ones?
Small and young firms grow faster: Gibrat’s law fails.
Firms that were smaller two years earlier grew faster in both employment and sales, and younger firms grew faster than older ones. Having a loan or credit line goes with faster growth (significant for employment, borderline for sales). Employment growth is highest among the youngest firms and lowest among the oldest.
Figure 19Does firm growth depend on size? (Gibrat’s law)Chart: annual employment growth by firm age, with 95% intervals.
Method and data
Method
Regress annualized employment growth (l1 vs l2) and sales growth (d2 vs n3) on log initial size and firm age; test whether the size coefficient is zero.
Questions used
l1l2d2n3b5k82h1d3ca4aa2a6a
Watch out
Only surviving firms are observed (survivor bias); sales growth is nominal.
A08Regression
Management practices and performance
Do better-managed firms have higher productivity?
Better management goes with higher productivity, at the margin.
A higher structured-management score is associated with higher sales per worker, though only at the edge of statistical significance. Firms where the owner is also the top manager score somewhat lower on management, at the margin of significance. Large firms have the highest average scores.
Figure 20Management practices and performanceChart: productivity difference for a one-standard-deviation higher management score (left) and correlates of the score (right).
Method and data
Method
Build a management score as the average of standardized r1–r11 answers (World Management Survey style), then regress log productivity on the score with controls.
Questions used
d2l1r1r2r3r4r5r6r7r8r9r10r11b3ab7
Watch out
Asked of a sub-sample; reverse causality (profitable firms adopt practices).
A09Descriptive
Firm life cycle: how much do firms grow after birth?
How large are mature firms compared with their size at start-up?
Canadian firms grow slowly over their life cycle.
The typical young firm is barely larger than at start-up. Even firms more than forty years old are, at the median, only a few times their initial size. Long-lived firms in professional services and manufacturing have grown the most; retail and hotels the least.
Figure 21Firm life cycle: how much do firms grow after birth?Chart: current employment relative to start-up employment by firm age (median and middle half).
Method and data
Method
Compare current employment with employment at start (b6) by age cohort; average growth multiple by sector (Hsieh–Klenow life-cycle comparison).
Questions used
l1b6b5b6ba4a
Watch out
Recall error for older firms’ start-up size.
3.3Finance
Access to credit, financing structure and payments.
A10Regression
Who is credit constrained?
Which firms are turned down or discouraged from borrowing?
Credit constraints are rare and not concentrated in one firm type.
Few firms are credit constrained, and no single trait predicts it strongly. Innovators and firms with audited accounts are slightly more likely to be constrained, but only at the margin. Among discouraged borrowers, unfavourable interest rates and expected rejection are the main reasons.
Figure 22Who is credit constrained?Chart: average marginal effects on the probability of being credit constrained.
Method and data
Method
Build a constrained indicator (rejected, or did not apply for reasons other than enough capital), then probit. Separate discouraged borrowers from rejected ones.
Questions used
k162k17k20a1l1b5k21b4b2bd3ch1a3a4aa2a6a
Watch out
Few constrained firms in Canada, so groups get small fast.
A11Advanced
Gender gap in access to finance
Do women-owned or women-led firms get credit on different terms?
No gender gap in access to credit.
Women-owned and women-led firms are as likely as others to have a loan or credit line. The decomposition finds no total gap, and nothing explained or unexplained. Female owners are no more likely to pledge personal assets or to be discouraged from applying.
Figure 23Gender gap in access to financeChart: Oaxaca–Blinder decomposition of the loan gap between women-led and men-led firms.
Method and data
Method
Probit for having a loan; OLS of collateral value relative to loan size (k15a/k11); Oaxaca–Blinder decomposition to split the gap into explained (firm traits) and unexplained parts.
Questions used
k82k14ak14dk15ak17b4b4ab7al1b5k21a4aa2a6a
Watch out
Collateral items are answered only by borrowers.
A12Regression
How firms finance working capital and investment
How do funding mixes differ by size and age?
Firms fund themselves mainly from internal cash.
Internal funds dominate working capital in every size class. The bank share rises with firm size and with audited accounts, while foreign-owned firms rely less on internal funds and slightly more on non-bank lenders.
Figure 24How firms finance working capital and investmentChart: average sources of working capital by firm size.
Method and data
Method
Fractional logit (Papke–Wooldridge) for each share; or describe the mix by size class.
Questions used
k3ak3bck3ek3fk5ak5bck5il1b5k21b2b
Watch out
Investment shares exist only for firms that bought fixed assets.
A13Regression
Digital payments and fintech adoption
Which firms take most of their payments electronically, and what does it cost them?
Online presence goes with digital payments.
Electronic payments are widespread in every sector. Firms with a website or social-media page receive a much larger share of payments electronically, and older firms a smaller share. Typical fees are low and money usually arrives within a day or two.
Figure 25Digital payments and fintech adoptionChart: average marginal effects on the share of payments received electronically.
Method and data
Method
Fractional logit for e-payment shares; describe fees and settlement days by sector.
Questions used
k33k38k342k36k35l1a4ac22ba3
Watch out
Fee questions are answered by e-payment users only.
3.4Labour & gender
Workforce, skills, training and gender gaps.
A14Regression
Female leadership and female employment
Do firms led or owned by women employ more women?
Women-led firms employ more women.
Firms with a female top manager have a markedly higher share of women in their full-time workforce; female ownership has a smaller positive link. Firm size and age make no difference.
Figure 26Female leadership and female employmentChart: average marginal effects on the female share of full-time workers.
Method and data
Method
Fractional logit of the female share of full-time workers (l5/l1) on female top manager and ownership, with sector fixed effects.
Questions used
l5l1b7ab4b4aa4aa2a6a
Watch out
Sorting: women managers may be concentrated in sectors that hire more women.
A15Advanced
Gender gap in firm performance
Do women-led firms differ in productivity and growth, and why?
A large productivity gap that firm traits do not explain.
Women-led firms report substantially lower sales per worker. Almost none of this gap is explained by size, age, manager experience, finance, trade, sector or region, so it is mostly unexplained. Sales per worker does not capture differences in activity within sectors, hours worked or profits, so this is not evidence of lower efficiency.
Figure 27Gender gap in firm performanceChart: Oaxaca–Blinder decomposition of the gap in log sales per worker.
Method and data
Method
Oaxaca–Blinder decomposition of log sales per worker between female- and male-led firms; check sector composition within 8 sectors.
Questions used
d2l1b7ab4b7b5k82a4aa6a
Watch out
Association only; sales per worker penalizes labour-intensive sectors.
A16Regression
Skills shortage as an obstacle
Which firms see an inadequately educated workforce as a severe obstacle?
Innovators feel the skills gap most.
Firms that innovate are significantly more likely to rate an inadequately educated workforce as a major obstacle. Workforce education, training, size and city size are not significant. Firms in Atlantic Canada, the Prairies and the North report skills problems more often than firms in Central Canada.
Figure 28Skills shortage as an obstacleChart: average marginal effects on the probability of rating skills a major or very severe obstacle (ordered probit).
Method and data
Method
Ordered probit of the 0–4 rating; compare with the share naming it the biggest obstacle (m1a).
Questions used
l30bl9bl10h1l1a3a4aa2a6a
Watch out
Perception data; firms in tight labour markets may rate it higher.
A17Regression
Who trains workers?
What explains formal training?
Large, R&D-active and foreign-owned firms train.
Foreign ownership, R&D and size are strongly linked with offering formal training. Among firms that train, women-led firms train a somewhat larger share of their female workers, at the margin of significance.
Figure 29Who trains workers?Chart: average marginal effects on the probability of offering formal training.
Method and data
Method
Probit for offering training; fractional models for the share of workers trained, including female workers (l12a).
Questions used
l10l11al11bl12al1h1h8b2bl30bl9b
Watch out
Training share items have sub-sample answers.
A18Descriptive
Labour flexibility: temporary workers, dismissal and severance
How costly is adjusting the workforce?
Temporary work is concentrated in seasonal sectors.
Hotels and restaurants, food, manufacturing and construction use temporary workers the most; retail and fabricated metal the least. Severance and dismissal times are modest and do not differ clearly across provinces. Firms that dismissed workers are more likely to see labour regulation as a major obstacle.
Figure 30Labour flexibility: temporary workers, dismissal and severanceChart: temporary workers relative to the full-time workforce by sector, with 95% intervals.
Method and data
Method
Describe weeks to dismiss and weeks of severance by region (provincial labour codes differ); relate to the labour-regulation obstacle with an ordered probit.
Questions used
l6l35l36l30al1a2a4al37
Watch out
Only firms that dismissed workers answer dismissal time.
3.5Trade
Exporting, trade partners and border costs.
A19Regression
Who exports, and how much?
What determines entry into exporting and export intensity?
Capabilities, not size, decide who exports.
Foreign ownership, R&D, quality certification and licensed foreign technology all raise the chance of exporting, while size does not once these are included. Among exporters, larger firms export a bigger share of their sales.
Figure 31Who exports, and how much?Chart: two-part model, export participation (left) and export share among exporters (right).
Method and data
Method
Two-part model or Heckman selection: probit for exporting (10%+ of sales), then a fractional model or OLS for the export share among exporters.
Questions used
d3bd3cl1b2bb8h8h1b5e6a4aa2a6a
Watch out
Needs an exclusion restriction for Heckman; two-part model avoids it.
A20Descriptive
Dependence on the United States
How concentrated are Canadian firms’ trade partners?
Most exporters depend on the U.S. and the land border.
In every province with enough exporters, most sell mainly to the United States, and the large majority of direct exporters ship through a land border. Atlantic exporters are the exception, with more diverse destinations, though their sample is small.
Figure 32Dependence on the United StatesChart: exporters whose main market is the U.S. by region (left) and main exit point (right).
Method and data
Method
Tabulate main destination and origin by sector and region; compare land, sea and air exit points.
Questions used
d31xd38xa4aa2d32d39
Watch out
Only the main partner is recorded, not the full mix.
A21Descriptive
Border frictions and trade costs
How long and how costly are customs and border procedures?
Imports face more border friction than exports.
Imported goods take longer than exports to clear customs and all border procedures, and compliance costs are higher for imports. Shipments through seaports and airports take longer than those crossing land borders.
Figure 33Border frictions and trade costsChart: days to clear customs and all border procedures (median and middle half of trading firms).
Method and data
Method
Weighted medians by exit/entry point and firm size; regressions of days on firm traits; tobit for cost shares at zero.
Questions used
d4ad33ad342d14ad40ad412d32d39l1a4a
Watch out
Small numbers of direct traders in some cells.
A22Regression
Imported inputs and productivity
Are firms using foreign inputs more productive?
Imported inputs go with higher productivity.
Firms that import directly and firms using more foreign inputs have significantly higher sales per worker. Exporting adds no further premium once importing is accounted for, and two-way traders show no extra gain.
Figure 34Imported inputs and productivityChart: differences in log sales per worker (OLS with sector and region controls).
Method and data
Method
OLS of log productivity on foreign-input share and direct importing; test two-way traders (import and export).
Questions used
d2l1d12bd13d3cb2ba4aa2a6a
Watch out
Selection into importing.
3.6Infrastructure & green
Utilities, digital adoption and climate.
A23Regression
Cost of power and internet disruptions
How much do outages cost firms, and where?
Power outages: similar across provinces, worse in the North and possibly in small towns.
Outage rates are similar in the four main regions, while the territories are a clear outlier. Firms in towns under 50,000 are more likely to experience outages and report somewhat larger losses, at the margin of significance.
Figure 35Cost of power and internet disruptionsChart: firms with at least one power outage by region, with 95% intervals.
Method and data
Method
Probit for experiencing outages by region and city size; tobit for losses as % of sales (many zeros).
Questions used
c6c7c8ac9ac39c40ac41aa2a3a4ac10
Watch out
Loss figures are firm estimates.
A24Regression
Digital adoption
Which firms are digitally connected, and does it pay?
Digital adoption is widespread.
Most firms have a website, file taxes online and handle most payments electronically. Larger and younger firms are more digital, and exporters slightly less. The digital index is not linked with sales growth.
Figure 36Digital adoptionChart: share of firms on each component of the digital index.
Method and data
Method
Build a digital index (website, e-payment share, broadband); regress on firm traits, then relate to sales growth.
Questions used
c22bk33c36l1b5b7a3a4aa2a6a
Watch out
Website ownership is near universal, so variation is limited.
A25Regression
Green practices: who adopts them?
What drives CO2 monitoring and energy-management measures?
Certification drives green practices.
Firms with an international quality certification are much more likely both to monitor CO₂ and to adopt energy-management measures, and both rise with size. The two practices go strongly together. Firms with a higher energy-cost share are less likely to monitor CO₂ and no more likely to manage energy, and past weather damage shows no link.
Figure 37Green practices: who adopts them?Chart: average marginal effects for CO₂ monitoring (left) and energy management (right).
Method and data
Method
Bivariate probit of the two practices; include energy cost share (n2b/d2) and past weather damage (ge3).
Questions used
ge7ge8dl1d3cb2bn2bd2ge3b8a4aa2a6a
Watch out
Self-reported adoption.
A26Descriptive
Climate exposure
Which firms suffered extreme-weather damage, and how much?
Climate damage is rare but uneven.
Few firms report assets damaged by extreme weather in the last three years. The territories and Atlantic Canada report it most often and British Columbia least. Where damage occurs, it is typically small relative to sales.
Figure 38Climate exposureChart: firms with weather damage by region, with 95% intervals.
Method and data
Method
Probit of damage by region and sector; describe damage relative to sales.
Questions used
ge3ge3aa2a3a4ag6b
Watch out
Rare event: few firms report damage.
3.7Government & institutions
Regulation, contracts, informality and perceptions.
A27Regression
Regulatory burden by firm size
Is regulation a heavier burden for small firms?
Tax compliance is a fixed cost that weighs on small firms.
Larger firms spend more total hours on taxes and are inspected more often, but per worker the burden falls sharply with size: small firms spend several times more tax hours per employee than large firms. Outsourcing tax work does not reduce reported hours.
Figure 39Regulatory burden by firm sizeChart: hours a year spent on taxes per full-time worker by firm size (median and middle half).
Method and data
Method
OLS of management time on regulation and hours on taxes, scaled per worker, on log size; check if outsourcing tax work (j34) changes the burden.
Questions used
j2j35j3l1j34j36a2a4aa6a
Watch out
Perception of time is approximate.
A28Regression
Who wins government contracts?
Which firms bid for and secure public contracts, and how slow is payment?
Larger and audited firms bid for public contracts.
Firm size and audited accounts are linked with bidding for or securing government contracts. Women-owned and women-led firms are not significantly different. Roughly half of contract holders find tender requirements difficult, and payment typically takes about a month.
Figure 40Who wins government contracts?Chart: average marginal effects on the probability of bidding for a government contract.
Method and data
Method
Probit for bidding or winning; describe payment days and tender difficulty by size.
Questions used
j6aj42j43j41l1b4b7ab2ba2a4aa6a
Watch out
Small number of contract holders.
A29Regression
Competition from informal firms
Which formal firms face unregistered competitors, and does it hurt them?
Informal competition is felt in food and professional services.
Competition from unregistered firms is reported most often in food, professional services, construction and hotels. Facing informal competitors has no significant link with innovation or sales growth.
Figure 41Competition from informal firmsChart: firms competing against informal firms by sector, with 95% intervals.
Method and data
Method
Probit of facing informal competition; then include it as a driver in innovation or growth regressions.
Questions used
e11e30a4aa3a2l1
Watch out
Perception-based.
A30Regression
Perceptions versus experience
Do firms that call electricity, crime or courts an obstacle actually experience more problems?
Electricity ratings track experience; corruption ratings do not.
Firms with more outages are much more likely to call electricity a major obstacle. Court ratings are related to having a commercial dispute only at the margin of significance. The small number of firms that report informal payments are significantly less likely to rate corruption a major obstacle.
Figure 42Perceptions versus experienceChart: effect of actual experience on rating each item a major obstacle (ordered probit).
Method and data
Method
Cross-tabulate ratings against objective experience (e.g. c30a against outages c6/c7); ordered probit of each rating on the matching experience variable.
Questions used
c30ai30h30j30fc6c7o1j7aj5
Watch out
Rating scale may be used differently across firms.
A31Advanced
Mapping the business environment
Which obstacles move together?
Sixteen obstacles reduce to four business-environment factors.
Factor analysis groups the obstacle ratings into tax and governance, markets and finance, infrastructure and order, and the labour market. Atlantic Canada scores worst on tax and governance, the territories and British Columbia on infrastructure and order, and the territories and the Prairies on the labour market.
Figure 43Mapping the business environmentChart: varimax-rotated loadings; darker cells mean the obstacle belongs more strongly to that factor.
Method and data
Method
Principal component or factor analysis of all 0–4 ratings to build indices (tax/regulation, infrastructure, labour); compare index scores across regions and sectors.
Ratings are ordinal; use polychoric correlations if possible.
3.8Ownership & location
Foreign and owner-managed firms, regions and cities.
A32Advanced
Foreign-owned firms in Canada
How do foreign-owned firms differ in productivity, exporting, training and certification?
Foreign-owned firms are more global and more formal, not more productive.
Compared with similar domestic firms, foreign-owned firms are far more likely to hold a quality certification and to export, and (by weighting, though not by nearest-neighbour matching) to train their workers. They are no more likely to do R&D or innovate, and their sales per worker is not significantly different.
Figure 44Foreign-owned firms in CanadaChart: difference from matched domestic firms (inverse-probability weighting, bootstrap 95% intervals).
Method and data
Method
Propensity-score matching or inverse-probability weighting of foreign (10%+) vs domestic firms on size, age, sector and region.
Questions used
d2d3cl10b8b2bl1b5a4aa2a6a
Watch out
Few foreign firms in small cells.
A33Regression
Owner-managed versus professionally managed firms
Do firms where the main owner is the top manager behave differently?
Owner-managers borrow more but manage less formally.
Firms run by their main owner are significantly more likely to have a loan or credit line, and score lower on structured management at the margin. They do not differ in productivity, innovation, training or growth.
Figure 45Owner-managed versus professionally managed firmsChart: owner-managed minus professionally managed firms, with 95% intervals.
Method and data
Method
Compare management, innovation and finance outcomes by b3a and legal form (b1) with controls.
Questions used
r1h1k82d2b3ab1b7l1
Watch out
Management items are sub-sample only.
A34Descriptive
Regional economies of Canada
How do firms in the Atlantic, Central, Prairies, BC and the North differ?
After sector and size controls, the provinces barely differ.
The raw regional differences seen earlier largely disappear once sector and size are held constant: the provinces do not differ significantly from Central Canada on exporting, innovation, credit or training. The one clear provincial difference is less R&D in the Prairies. The territories differ more, but their sample is very small.
Figure 46Regional economies of CanadaChart: difference from Central Canada after sector and size controls.
Method and data
Method
Weighted indicator tables by region, then regressions with region fixed effects to see which differences survive sector and size controls.
Questions used
d2d3ch1k82l30bc6a2a3a4aa6a
Watch out
The North has very few firms; pool or report with a warning.
A35Regression
Urban premium
Are firms in big cities more productive and innovative than small-town firms?
Big-city firms innovate more but are not more productive.
Firms in cities over one million people are more likely to innovate than firms in small towns. There is no significant urban premium in sales per worker or in exporting.
Figure 47Urban premiumChart: difference in the probability of innovating relative to towns under 50,000.
Method and data
Method
OLS of log productivity on city-size categories (a3) with sector and region controls.
Questions used
d2l1h1a3a3ca2a4a
Watch out
City size is grouped into four bands.
3.9Beyond Canada
Comparisons with other economies.
A36Advanced
Canada versus the U.S. and U.K.
How do Canadian firms compare with firms in similar economies?
Not run yet.
This comparison needs the U.S. 2024 and U.K. 2024 Enterprise Survey datasets, available through the same World Bank portal.
Method and data
Method
Pool Canada with the U.S. 2024 and U.K. 2024 Enterprise Surveys (same questionnaire); regressions with a country dummy and sector-size controls.
Questions used
h1d3ck82l10b7aa6aa4a
Watch out
Requires access to the other countries’ data under your account.
Part 4
Data and method
4.1Data, weights and definitions
Data. World Bank Enterprise Survey, Canada 2024 (fieldwork April 2024 – April 2025): formal private firms with five or more employees in manufacturing and services, stratified by sector, size and region.
Weights. All figures use the survey’s median-eligibility weights, so they describe the population of eligible Canadian firms; variance estimates account for the stratified design. The quantile regressions in analysis A05 are unweighted.
Groups. Size: small 5–19, medium 20–99, large 100+ employees. Regions: Atlantic, Central (Ontario and Quebec), Prairies, British Columbia, and the territories, whose sample is very small.
Definitions. Exporter: direct plus indirect exports at least 10% of sales. Major obstacle: rated “major” or “very severe”. “Don’t know” and refusals are excluded.
Interpretation. One cross-section of self-reported answers; results are associations. With many models, a few borderline results may arise by chance, so the clearly significant ones deserve the most weight.
The firm-level data are licensed by the World Bank and are not reproduced or redistributed on this page. Figures show weighted aggregate estimates only. Researchers can request the data from the Enterprise Surveys Data Portal.
