The Cost of Language Barriers in Multinationals

Most multinationals know language barriers are a nuisance. Few have measured what they cost. A study published in the Review of Economics and Statistics did, inside 27 foreign-owned companies, and then ran an experiment to see what happens when the barrier is lowered.

The paper, "Language Barriers in Multinationals and Knowledge Transfers" is by Louise Guillouët, Amit K. Khandelwal, Rocco Macchiavello, Madhav Malhotra and Matthieu Teachout, researchers affiliated with Columbia, Yale and the London School of Economics. Yale's Economic Growth Center lists it among its research. It is set in Myanmar, but the problem it describes will be familiar to any company where head office works in one language and the local team in another.

What it found:

• A third of meeting time was lost. Local managers said 33.8% of the time they spent in meetings with foreign managers was lost to the language barrier, and that they understood about 60% of what was said.
• Translation was a weak fix. Google Translate raised comprehension by only 2.5 points. A human translator raised it to 84%, but cost about six times a local manager's daily wage, and only half the firms used one.
• A 48-hour course changed behaviour. Managers randomly offered a free business English course talked more with their foreign bosses, became more involved in managing staff, and in a test task ran a team that finished 15% faster with no more mistakes.
• The skills transferred. More than a year later, the trained managers reported learning more soft skills on the job, such as problem solving, negotiation and confidence.
• Employers underinvest. Only two of the participating firms had paid for any formal English training for their local managers, although 90% of the firms invited took part. The authors argue the gap is structural, not a one-off oversight.

The setting: three layers and no shared language

The researchers studied multinationals in Thilawa, Myanmar's main special economic zone, where the largest investors are Japanese, followed by Singaporean, Thai, Chinese, European and American, and Korean companies. Each firm had the same three-layer structure:

• Foreign managers, about 4.5 per firm, posted from head office for around two years. They set direction and connect the subsidiary to headquarters.
• Domestic managers, about 13 per firm, local hires who run meetings, handle staff, deal with suppliers and coordinate across departments.
• Production workers, supervised by the domestic managers in Burmese.

The working language between the first two layers was English, which was the native language of neither. Almost none of the foreign managers spoke Burmese, and very few local managers spoke Japanese, Korean or Chinese.

The middle layer is where it matters. Domestic managers said that learning new skills was the main reason they joined a multinational (58% gave that answer, against 10% for higher wages), and management skills above all. They spent about 1.1 hours a day with foreign managers. But when the course provider tested a sample of them, 85.4% were at A2 or below on the CEFR scale, meaning they could handle simple, familiar exchanges but not much beyond.

What the language barrier cost

The paper's surveys put numbers on what that gap does in practice.

Measure	Result
Meeting time with foreign managers lost to the language barrier	33.8%
Share of the conversation understood	60.4%
Share understood with Google Translate	62.9%
Share understood with a translator present	84.0%
Firms that hired translators	About half
Cost of a part-time translator	About US$100 a day, against a local manager's US$15.70

Translators were the only fix that worked, and they worked badly as a business. They are expensive, and a translated meeting covers about half the content of a direct one, because everything is said twice. Firms that used them hired them for about 1.5 hours a day, and the translator followed the foreign manager rather than the local team. The authors estimate that one translator per foreign manager would cost the typical firm about US$450 a day, more than twice the daily wage bill of its entire middle-management layer.

The people involved described it more simply:

"He's not an English native speaker and we are not native speaker either. So, although we try our best, there are misunderstanding frequently. Sometimes, we don't know what he wants." — Domestic manager, Japanese firm

"I asked FMs to repeat what they said for more than two times, and was afraid to speak with them." — Domestic manager, Korean firm

"I told the staff to handle products in certain way but they didn't really understand it and did it differently." — Foreign manager, Japanese firm

Note the word afraid. Several managers described the barrier as fear of speaking up as much as missing vocabulary. That matters for what kind of training helps.

The experiment: a free 48-hour business English course

Correlations only go so far: managers with better English might simply be more ambitious. So the researchers ran a randomized trial. In 2019 they recruited 298 domestic managers across 27 multinationals, and within each firm randomly offered half of them a free business English course from a Yangon language school.

The course was ordinary on purpose: 48 hours over three months, two two-hour sessions a week, after work or on weekends, taught by native English speakers. It was the provider's standard business English course, with daily-life scenarios swapped for workplace ones.

Attendance was typical of after-hours training. Managers attended 56% of the sessions on average, and 45% attended at least three-quarters of them. English scores still rose by 0.15 standard deviations among those offered the course, and by about 0.26 among those who attended.

What changed when the barrier came down

More conversation with head office, not with everyone. Managers offered the course reported talking with their foreign manager more often, about 34 more minutes a week. They did not talk more with their local colleagues. The extra communication went specifically to the relationship the language barrier had been blocking, which suggests it was not just a general boost in confidence or a wish to please the researchers.

More involvement in management. They reported being more involved in managing employees: setting targets for subordinates, promoting good performers and hiring.

Better relaying of instructions. The researchers built a management simulation. An English-speaking "foreign manager" gave a domestic manager instructions for a packaging task, which the domestic manager then had to get two Burmese-only "workers" to carry out. Trained managers spent 19% more time in conversation with the "foreign manager", mostly because they asked more questions, and their teams completed the task 15% faster, with no more mistakes. When the instructions were given in Burmese instead, the two groups performed the same, so it was the language that made the difference, not extra effort.

Skills that lasted. In a follow-up more than a year later, trained managers reported learning more soft skills during their time at the company: problem solving, negotiation, confidence, professionalism and customer relations. They also applied to more jobs, though wages and turnover did not differ. That follow-up was disrupted by the pandemic and by the February 2021 military coup, both of which cut face-to-face contact and job opportunities.

The job market values those conversations

To check whether the time spent with foreign managers was actually worth something, the researchers asked 51 HR managers at domestic Myanmar firms to rate fictional résumés and interview answers for a general management job, varying one detail at a time.

• Multinational experience earned a 10.2% higher salary offer.
• An applicant who said they interacted frequently with their foreign manager earned a 9.9% premium on top, about as much as one who listed a concrete hard skill (frequent use of Microsoft Office).

An applicant who said they had been picked to work closely with their foreign manager, but never got to, earned almost nothing extra. Employers were paying for the interaction itself, not for being the kind of person chosen for it. That is the paper's case that what flows through those conversations is real management know-how, valued outside the firm.

Why companies underinvest in language training

If the payoff is that clear, why had only two of the participating firms paid for any formal English training for their local managers, when 90% of the firms invited took part in the study?

The authors' answer is that communication inside a firm can't be written into a contract. A foreign manager decides how much to explain based on getting today's work done, not on how much the local manager learns. So the firm captures only part of the value of better communication, the employee captures another part, and neither invests enough. The authors use a model to show this is a reason for public support of language training, not only a case for employers to spend more.

They also list more ordinary reasons a firm might hold back: overestimating the chance that trained staff will leave, underestimating the returns, or simply being too busy to find and organize a program.

What this means for employers in Canada

The study did not look at Canada, and Myanmar's factories in 2019 are not a Montreal head office in 2026. But the mechanism carries over to any company where the people setting direction and the people running the work use different first languages:

• A multinational's Quebec subsidiary, where the parent company works in English and the Quebec team works in French, which the Charter of the French Language gives workers the right to do.
• An Ontario or national organization serving francophone clients, where bilingual frontline staff report to managers who mostly work in English.
• A Quebec company working with English-speaking head offices, suppliers or clients, where the local team's English becomes the bottleneck.

Four lessons translate directly.

1. Measure the time, not only the misunderstandings. The study's most striking number is the third of meeting time lost. To size the problem in your own team, multiply the hours your managers spend in meetings held in their second language by the share they say is lost, then by their loaded hourly cost. It won't be precise, but it turns a vague frustration into a budget line.

2. Look at the middle layer. The barrier that mattered was between the people who set direction and the people who relay it. Training front-line staff while leaving their managers unable to talk comfortably with head office misses the link that carries know-how down.

3. Translation tools reduce the problem, but don't remove it. In this study, Google Translate barely helped, partly because it had to go from Burmese to Japanese through English. Machine translation has improved a great deal since 2019, and it is far stronger between French and English. But a translated meeting still loses the follow-up question, the quick aside and the confidence to push back, which is where management know-how flows.

4. Train for speaking under pressure, not only vocabulary. The managers in this study described being afraid to speak. Even a standard course that participants attended about half the time measurably changed their behaviour. Practice that builds confidence in the actual situations, such as the weekly meeting, the call with head office or the difficult conversation, targets the specific barrier the study found.

Quebec employers have some of the policy support the authors call for. Bill 90 requires employers with a payroll above $2 million to invest 1% of it in training, and language training can count toward it, depending on how it is delivered and documented (our guide covers what qualifies). Francisation Québec offers workplace French courses, free for eligible companies. And under Bill 96, companies with 25 or more employees must register with the OQLF and analyse how French is used, which often brings the language gaps between head office and the local team to light.

Limits of the study

The paper is careful about what it shows, and so should we be:

• One country, one kind of firm. Factories in a special economic zone, with English as a second language on both sides. Results may differ where one side is a native speaker, or in office work.
• A modest sample. 298 managers in 27 firms is large for a field experiment, but small enough that several effects are imprecise.
• Mostly self-reported outcomes. Communication and skills came from surveys. The management simulation is an objective measure, but it is a simulation.
• No firm-level profits. The study shows more communication and transferable skills, not a change in the companies' bottom line.
• A disrupted follow-up. COVID-19 and the 2021 coup hit exactly the period when long-term effects would have shown up.

Where Conversaflex fits

Conversaflex is an AI conversation practice platform for French, English and Spanish at work. Employees speak out loud with an AI partner in scenarios built around their role, such as the meetings, calls and presentations they actually face, starting from a CEFR placement test, with pronunciation and grammar feedback as they speak. Sessions typically run ten to twenty minutes, at any hour, so practice does not depend on everyone attending the same after-hours class. HR sees practice days, hours and conversations for each employee.

It is not a teacher and is not meant to replace one. It is built for the problem this study describes: people who know more of the language than they can comfortably use in a meeting with their boss. Our corporate language training page explains how a program is rolled out, the page on English language training for employees covers English for French-speaking teams, and you can try a limited version yourself.

Frequently asked questions

What is the cost of language barriers in multinationals?

In the study of 27 multinationals in Myanmar, local managers said 33.8% of their meeting time with foreign managers was lost to the language barrier, and that they understood about 60% of those conversations. The researchers also found the barrier blocked management know-how from moving from foreign to local managers, a cost that does not show up on any invoice.

Does language training improve communication at work?

In this study, yes. Managers randomly offered a free 48-hour business English course talked with their foreign managers about 34 more minutes a week, became more involved in managing staff, and led a team that completed a test task 15% faster with no more mistakes. They did not talk more with local colleagues, so the effect was specific to the relationship the language barrier had been blocking.

Can translators or translation tools replace language training?

Not fully, according to this study. A human translator raised comprehension to 84%, but cost about six times a local manager's daily wage and halved the content each meeting could cover. Google Translate raised comprehension by only 2.5 points in that setting. Machine translation is stronger between French and English today, but it still slows down the quick exchanges where know-how is passed on.

Why don't companies invest more in language training?

The authors argue that because communication between managers can't be written into a contract, managers explain only what today's work requires, and neither the firm nor the employee captures the full value of better communication. Firms may also overestimate turnover, underestimate the returns, or lack the time to organize training. Only two of the firms in the study had paid for any formal English training for their local managers.

Who wrote the study and where was it published?

Louise Guillouët, Amit K. Khandelwal, Rocco Macchiavello, Madhav Malhotra and Matthieu Teachout, researchers affiliated with Columbia, Yale and the London School of Economics. It was published in the Review of Economics and Statistics (vol. 108, no. 5, 2026, pp. 1270–1288). An earlier version circulated as NBER Working Paper 28807.

Source: Guillouët, L., Khandelwal, A. K., Macchiavello, R., Malhotra, M. and Teachout, M. (2026). "Language Barriers in Multinationals and Knowledge Transfers." Review of Economics and Statistics, 108(5), 1270–1288. doi.org/10.1162/rest_a_01487. Figures are from the authors' accepted manuscript. Amounts are in US dollars.