*The same model costs about a million rubles in China and almost twice that in Russia. Taxes and duties are not the whole explanation: the price also reflects which industries the country does not have. We follow sixty years in the history of eleven countries to work out why a production line of one's own still does not make a country an automotive power.*

## Two prices for one car

The Moskvich 3 is a close relative of the Chinese JAC JS4, which is sold on the Chinese home market as the Sehol X4. The two cars share the same body and the same platform. In China such a car cost 75.8–98.8 thousand yuan — roughly 0.95–1.25 million rubles at the current exchange rate. In Russia the same car is sold at an official list price of 1.95–2.24 million, and dealer promotions in the autumn start from 1.99 million (2026). That makes a price gap of almost two to one.

Why does the same model cost so differently in two countries? The simplest answer is that the sellers mark it up too much. But the explanation "the gap is the price of the missing production chain" is not quite right either. Three main factors affect the price. **The first factor is taxes and duties.** The vehicle recycling fee is levied both when a finished car is imported and when a car is built in Russia. A manufacturer can, however, receive compensation in the form of a subsidy if localization requirements are met. As the Moskvich 3 will show, not everyone reaches the required threshold. For cars with an engine of one to two litres the rate has risen since October 2024 from 300.6 to 556 thousand rubles and has kept being indexed since. VAT is added on top. These payments alone explain a significant part of the difference. **The second factor is logistics and the dealer margin. The third is the state of industry itself:** how developed the local supplier network is. There may be no local supplier making springs, controllers or cast parts. There may also be no engineering school capable of combining such components into a product of its own. After 2022 many links in that chain had to be found anew. Someone ultimately pays for every missing link. The price tag does not separate these causes: the buyer sees one final sum. And the state of industry has to be assessed by other signs — localization scores, the origin and age of the platform, the makeup of the supplier base. The rest of this analysis is built on that.

The Volga brand shows why a price tag alone is not enough. The brand has returned: since 19 June 2026 it sells the K40 crossover from 2.75 million, the C50 sedan from 2.9 million and the K50 flagship from 4.2 million rubles. According to media reports and certification databases, the platforms underneath are Geely's: the K50 repeats the Monjaro, the K40 is built on the Atlas, the C50 on the Preface. Brand representatives deny cooperation with Geely. That in itself does not prove the opposite, but the similarity of the models remains a matter of discussion. So: the Geely cars of the same name cost more in Russia than their twins. The Volga does not yet have a production chain of its own, yet that does not mean its price must necessarily be higher than the price of the original model. Price is affected by more than one factor: local assembly and the chosen positioning can even make a car cheaper than the model it derives from. That is why the absence of a production chain of one's own has to be proven by looking at production, not by comparing prices.

Next we will look at what the car industry can tell us about a country's economy. That will take sixty years of history and eleven countries.

![Two prices for one car: what a missing production chain costs](/upload/upload-why-some-countries-build-cars-fig-2-two-prices-one-car.svg)

Fig. 1 · paired bar chart. "The same car": the JAC JS4 (Sehol X4) in China against the Moskvich 3 in Russia, with the gap broken down by layer (recycling fee, VAT, remainder); counter-example: the Volga K40/C50/K50 against the Geely models of the same name in Russia. 2026 price lists, yuan converted at the year's rate. Sources: JAC/Sehol, Geely and Moskvich price lists, dealer price lists.

## Why the car in particular

Among mass-market goods the car is one of the hardest to make. A car has twenty to thirty thousand parts, and its production rests on several tiers of suppliers: from metallurgy to electronics and software. Mass output — millions of cars a year — has to be combined with precision: tens of thousands of operations must run without defects day after day.

Aircraft can be technically more complex than a car, but they are built in orders of magnitude smaller numbers. Their production chains are usually organised differently, and governments play a large role among the customers. Smartphones are produced in enormous volumes, but manufacturers buy a significant share of their components from specialised suppliers. The car is especially telling: it is a complex product of mass production that millions of people buy. That is why the capabilities of an industry are visible not only to specialists — they show up in the price, in the quality of the car and in the availability of spare parts. The Harvard Atlas of Economic Complexity, linked to the work of Ricardo Hausmann and César Hidalgo, is built on a similar logic: it assesses which knowledge and production capabilities are needed to make different goods. A car demands a combination of many competences, so the ability to create one independently usually says something about other industries as well.

The car industry can therefore serve as a quick and vivid test of a country's industrial potential. Not because a car matters more than an aircraft or a processor, but because weaknesses are hard to hide in car production. Natural resources alone, a single strong laboratory or assembly from ready-made kits are not enough. What matters is whether the economy can sustain the entire production chain.

That test, however, is easy to misread. Output volume is not the indicator: Mexico has held a place in the world's top ten producers for decades, seventh in 2025 with 4.1 million cars, and has not become an automotive power. The mere existence of an assembly line proves nothing either: if a plant only bolts together a kit brought in from abroad, that is not the same as a developed car industry. Two things have to be assessed: the production chain — whether the country has suppliers and production of its own — and the engineering school, meaning the ability to develop new models and technologies from one generation to the next.

That is the main thesis of the article. The car industry is the best quick test of a country's potential if you measure the chain and the school rather than the volume. Over the past sixty years only three countries have travelled that road from a relatively low base: Japan, South Korea and China. In all three cases the rise coincided with a major change in technology or the opening of a new market.

The thesis has three parts, and each has to be argued. **It is possible to climb the steps.** The examples are Japan of the 1960s, South Korea of the 1970s and 1980s, and China of the 1990s through the 2010s. **But there is also a ceiling.** A country can remain an assembly site for decades. Mexico and Spain have produced millions of cars for decades without ever creating a school of their own; Brazil has a school, but in exactly one line, ethanol engines, and beyond that line it went no further. **A leap needs a suitable window of opportunity.** Japan was helped by the oil crises, South Korea by a coherent state industrial policy, China by the shift to electric traction and access to a large market.

## Seven indicators: how to measure

Before turning to history, let us define the seven indicators by which the production chain and the engineering school can be assessed. At the end of the article they will come together into a checklist for a first-pass assessment of any country.

**The first indicator is depth of localization.** A claim of "95% localization" says little if only the list of parts is counted. A finished assembly could have been bought abroad and then formally counted as a local component after being assembled inside the country. Localization is therefore better assessed by added value: what share of a car's price is actually created inside the country. A warning sign is high localization on paper combined with a production line stopped by the shortage of a single microchip.

**The second indicator is platforms of one's own.** A platform is the basis of a car: the load-bearing body structure, the floor pan and the overall layout. The age of the platform and who owns it say more about an engineering school than the ceremonial launch of yet another production line. A platform of one's own is a sign of a working engineering school. If a licence half a century old is passed from generation to generation, that may mean independent development has stopped.

**The third indicator is tier-2 and tier-3 suppliers.** Does the country have companies of the level of Denso, CATL or Mobis — makers of components the buyer never notices but without which a car cannot be assembled? If key components come from local subsidiaries of Bosch, ZF and Denso, that still does not mean a supplier base of its own has formed in the country, however many cars roll off the line.

**The fourth indicator is engineering personnel.** The number of technical graduates, the outflow abroad, the average age of design engineers. This indicator reflects the current state of the workforce.

**The fifth indicator is price against level.** The price of a car in dollars is set against its technical level. If the car gets more expensive while its platform hardly changes, that may point to lagging engineering development or to a high cost of production.

**The sixth indicator is where exports go.** Sales in the United States, Western Europe and Japan are a serious test for a manufacturer: the car has to meet the requirements of buyers and regulators. If cars can only be sold in markets with looser requirements, that is a reason to assess the level of the engineering school more cautiously.

**The seventh indicator is where engineers come from.** Here it matters whether a country trains its own specialists, attracts them from abroad, acquires knowledge through joint ventures or loses personnel. The fourth indicator describes the state of the workforce, while the seventh describes the system that trains and retains it. This indicator changes slowly, but to a large extent it determines the others: a personnel system cannot be rebuilt in a single year.

**T-3. Seven indicators: how to measure**

| No. | Indicator           | How to measure                                            | Red flag                                                  |
| --- | ------------------- | --------------------------------------------------------- | --------------------------------------------------------- |
| 1   | Localization        | Added value inside the country, not the nomenclature      | "96%" on paper, a line stopped by one missing part        |
| 2   | Platforms           | Age and origin of the chassis, who owns the documentation | A licence from the 1970s passed down through generations  |
| 3   | Tier 2–3 suppliers  | Whether the country has component makers of world level   | Key suppliers are subsidiaries of foreign groups          |
| 4   | Personnel           | Engineering graduates, outflow, age of design engineers   | The best leave the country, the design bureau grows older |
| 5   | Price against level | Price in dollars against the level of the platform        | The price rises, the platform stands still                |
| 6   | Where exports go    | Share of sales in markets with high requirements          | Only markets with looser requirements                     |
| 7   | Personnel model     | Grows / recruits / acquires / loses                       | A school exists, but there is no demand for it            |

*Caption: the author's methodology; data sources by row: OICA, UNESCO UIS, OECD, national associations (the full summary is in the working plan, section 8).*

The seventh indicator deserves a separate discussion: engineers do not fall from the sky.

## Grow, recruit, acquire: where engineers come from

An engineering school is not a building and not a university faculty. It exists where a country knows how to train people capable of developing new solutions and to pass the experience on to the next generation. That is the sense in which the word "school" is used here. The models differ from country to country, but the main approaches repeat.

**Germany and Japan grow their own systematically.** German dual training, where half the theory is taught at school and half the practice happens at the plant, keeps roughly half of young people after school; more than 432 thousand firms had apprentices in 2019, every fifth company in the country. Japan teaches inside companies: you come in as a schoolboy and leave as a master. **The French build an elite with public money:** the polytechniques and the grandes écoles. **The Chinese press with scale;** nobody knows how many engineers China graduates a year: estimates range from 1.5 million engineering bachelor's degrees to 3.5 million STEM graduates, and NCSES adds a third figure, 2.0 million first university degrees in science and engineering for 2020. **The United States recruits:** a magnet for other people's talent, and in the 2022/23 academic year India overtook China as the source of foreign students for the first time since 2009, with 268.9 thousand people. **They acquire through joint ventures:** that is how Korea rose in the 1970s and China in the 1990s; the Korean Pony of 1975 received its engine and transmission from Mitsubishi and its body from ItalDesign. **They lose:** Russia, with its 520 thousand STEM graduates a year, a good mathematics school and a constant outflow of people.

When comparing these figures it is important to keep in mind the difference in counting methods. "Engineer", "STEM" and "engineering, manufacturing and construction" mean three different categories; older comparisons counted three-year programmes and vocational school graduates into the Chinese and Indian numbers. An honest count requires one category from UNESCO UIS or OECD and a count per million inhabitants rather than per country. A count per million inhabitants shows how developed engineering training is relative to the population, not simply how large the country is.

The state of an engineering school is shown not only by the number of graduates but by results: which solutions it creates and whether they can be turned into products of its own in series.

**T-4. The school and the fate of an idea**

| Country | Idea                                                                                                                                 | What came of it                                                                                                   |
| ------- | ------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------- |
| Britain | Mini (1959): transverse engine plus front-wheel drive                                                                                | The layout will conquer the world, but not for Britain: the country will lose its mass industry (British Leyland) |
| Sweden  | Volvo three-point seat belt (1959, Nils Bohlin); patent opened to all, US No. 3,043,625                                              | Leadership in an idea at small volumes: the belt became common property                                           |
| France  | Citroën DS (1955), hydropneumatics                                           | A revolution in suspension; other companies answered with simplified copies                                       |
| Italy   | Common rail: invented at the Fiat and Magneti Marelli centre, brought to series by Bosch; first series car Alfa Romeo 156 JTD (1997) | They could invent, they could not hold on                                                                         |
| Germany | ABS by Bosch (1978)                                                                                                                  | A school plus a supplier network: the idea stayed at home                                                         |
| Japan   | Prius (1997), the first mass hybrid                                                                                                  | A technology window, used first                                                                                   |
| Russia  | Lada B/C (2014–2015), a platform of its own                                                                                          | The school lived to the last, the analysis is in the Russia case                                                  |

*Caption: plant and museum histories, patent databases; the summary of sources is in the plan, section 8.*

For the seventh indicator the warning sign is the same as before: engineers are trained, but there is not enough demand for their work inside the country. For the fourth, personnel, the red flag has to be named more precisely. It is not the graduate who went into IT after an engineering degree: a person goes where they are wanted, and blaming them for reading the market is strange. Two other signals matter here. The first is the outflow of specialists: engineers who have left are no longer counted in national employment statistics. In Russia in 2022, 71.5% of engineering graduates worked in their specialty, and even that figure does not see those who packed a suitcase: one country invested in training a specialist, and another gets the benefit of their work. The second signal is the age of the workforce. If the average engineer in a design bureau is over fifty and young specialists are few, that may mean the school is not renewing itself but resting on the experience of the older generation. We will see both flags again in the Russia case. Without industry of its own, an engineering school risks becoming not a source of development but a channel through which trained people drain away.

Three rows of the table — Britain, Sweden and France — will return below, at the very end of the 1950s. Now let us turn to history: how the world's automotive leader changed and what those changes say about the economies of the countries that lost first place.

## The ribbon of leaders: how first place changed hands

Before looking at the figures it matters to understand what exactly they show. Output by country mixes two different questions: **where a car is assembled** and **whose car it is**. An American Honda plant counts toward US output; a German BMW plant in Spartanburg ships cars into German exports. The ribbon below measures output, that is, the geography of plants. It does not show the school and the chain: we will apply those to each leader separately. The main rule is simple: volume shows the scale of production but by itself does not explain how developed the industry is. And one framing note: the ribbon stretches from 1945, but the race itself covers the last sixty years; up to 1965 comes the prologue and the line-up.

### 1945–1965: the line-up

The car industry ended the Second World War with a predictable result: the plants of Europe and Japan lay in ruins, the American ones stood in line and worked for the front. In 1950 the United States accounted for about 8 of the 10.6 million cars of world output, roughly three quarters. The share looks like a monopoly; and it was one: for fifteen years the world bought American because there was nobody else to buy from.

Those two decades are not part of our race but its line-up. Yet that is exactly when all the future heroes step onto the start. Japan makes two bets that will define half a century for it — exports and quality: the Toyota production system will be born out of a poverty of resources. The USSR buys a licence and a plant from Fiat, and in 1970 VAZ appears. Europe rebuilds and, as we will see, manages to produce more ideas than anyone.

### The late 1950s: Europe at its peak of ideas

A pause between changes of leader. Four years, four ideas, three countries.

1954: Citroën fits hydropneumatics to a production model for the first time — the rear axle of the Traction Avant 15-H. 1955: the series Citroën DS with hydropneumatics makes a revolution in suspension that lasts decades, and the world answers with simplified copies. 1959: the British Mini brings the transverse engine and front-wheel drive, a layout that will conquer every city in the world. The same 1959: the Swede Nils Bohlin patents the three-point seat belt for Volvo, and the company opens the patent to all. Sweden thereby made an important technical solution available to the whole world.

In those years the United States led in output volume, but the innovative solutions listed above appeared in other countries. That distinction matters: leadership in the number of cars does not always mean leadership in engineering. A country can assemble a lot and invent little. And the other way round.

(An echo forward: the same years were the heyday of Italian, French and British cinema. More on this coincidence at the end, in the conclusion.)

### 1965–1979: oil changes the rules

The period under review begins, and soon comes the first big break. Formally first place still belongs to the United States, but the rules are changed not by the leader but by oil. Two shocks, in 1973 and 1979, make economy rather than power and size the main quality of a car overnight. Demand switches to small reliable cars, and that is a field on which Japan already stood ready.

In 1974 Japan overtakes West Germany and becomes the world's largest car exporter. The reason is clear: fuel becoming more expensive changed demand, and Japanese manufacturers already knew how to build the cars that were now especially sought after. While American plants moved from dreadnoughts to compact cars with great difficulty, Japanese ones grew by 10–15% a year.

In the same decade Korea starts, a country that built no cars of its own in 1970 at all: a borrowed Cortina had been assembled for Hyundai since 1968, but there was no model of its own. The choice was naive and precise: assemble a model of one's own from other people's parts, take the engine and transmission from Mitsubishi and the body from ItalDesign. That is how the Pony is born in 1975. The Pony did not yet mean full independence, but it set the direction: use foreign technology as a step towards development of one's own.

### 1980–1994: the change of leader and its price

1980 brings the first real change of first place since the post-war years. Japanese output grows to 11.04 million cars, up 14.6% in a year; American output falls by more than 30%, to 8.01 million. In 1980 more than half of Japanese output went to export, up to 54%; demand for passenger cars in the United States contracted by 16% in the same year. A recession in America met Japan's bet on quality, and the leader changed because of two causes together, not one.

It is important to understand the mechanism itself. Japan moved ahead not only thanks to quality: the conditions of the market changed, and it was small, economical and reliable cars that came out ahead. By the time of the change Japanese industry was already prepared to take advantage of the new opportunities.

Next comes the most instructive page for the theme of this article. America answers the Japanese onslaught not with engineering but with a trade rule: from May 1981 "voluntary" export restraints (VER) fix Japanese shipments at 1.68 million cars a year; later the ceiling is raised to 1.85 million (1984) and 2.30 million (1985), and the programme is closed in 1994.

One detail matters here: cars assembled in the United States did not fall under the quota. And the Japanese do the obvious thing: they move. November 1982: Honda opens a line in Marysville, Ohio. 1984: GM and Toyota launch NUMMI together, a single plant where two schools learn from each other. By 1990 the quotas had lost their meaning: production inside the United States had displaced direct exports.

Two conclusions follow from this story. First, trade restrictions do not necessarily stop a competitor — they may force it to move production. Second, from that moment the count of "where it is assembled" and the count of "whose brand it is" diverge for good. It is precisely this crack that will prepare the next turn of the ribbon.

### 1994–2005: the United States takes first place back

In 1994 America is first again. There is historical irony in it: the VER programme was closed in the same 1994, and the ceiling had been lowered to 1.65 million back in 1992: the quotas stopped being needed before they were officially buried. By that point the restrictions decided nothing, and first place returned to the country that had introduced them.

What brought the United States back to the top? It was not so much a victory of American engineering as the strength of the American market and the dollar. By the mid-1990s Honda, Toyota and Nissan were assembling millions of cars in America, and every one of them went into US output. The United States regained first place not only through its own plants but because it began to attract other companies' production: Detroit never reforged its compact-car school, but it pulled someone else's onto its own soil, to the buyer and to the dollar. Add Japan's decline after the bubble burst in 1990. Add NAFTA (1994), under which Mexico slots into the American chain as a cheap workshop.

There is a third version too: "the collapse of the USSR opened markets and freed America's hands". It sounds fine, but it has few verifiable channels, and in the ribbon it remains the context of "coincided in time" rather than a cause.

A natural question arises: who is the leader — the country where the cars are assembled, or the country that owns the brand? The honest answer: by output it is the United States; by schools there are still two. The ability to attract production is also an advantage: a large and rich market creates jobs, brings taxes and lets even foreign engineering schools operate. But the ribbon measures plants. Schools live longer than trade rules.

### 2006–2008: Japan's short return

In 2006 Japan briefly takes first place back: US output contracted amid a crisis. This time the rules did not change: it was simply demand. Two years later the world crisis will hit everyone indiscriminately and prepare the largest turn of the ribbon.

### 2009–2021: China moves into the lead

2009: China overtakes Japan and becomes the world's largest producer with a result of 13.8 million cars. And since then nobody has overtaken it.

China's rise is explained by a combination of three factors. The first is joining the WTO in 2001: foreign companies gained access to the market, and joint ventures with local partners became an important way of working. The second is technology transfer and accumulated experience: engineers learned inside such enterprises. The third is the scale of domestic demand: a fast-growing economy became the world's largest car market. In parallel an engineering school of its own developed: the share of local suppliers grew, national brands appeared, and by the 2020s China had created battery makers of world level. The causes: trade policy plus support policy.

### 2022–2026: China becomes the top exporter

This is the last period the article covers. China's exports: 3.11 million cars in 2022, 4.91 million in 2023, 5.86 million in 2024. In 2023 China overtook Japan and became the world's top exporter, 4.91 million against Japan's 4.42 million according to national associations. In 2025 passenger car exports reached 6.04 million.

An important caveat: data on all motor vehicles and data on passenger cars alone differ, and so do the counts of the industry association CAAM and of customs. The difference can reach half a million cars. Here a single series is used: CAAM data on all motor vehicles.

There are two causes. The first is the shift to a new technology: 12.9 million electric cars were produced in 2024, up a third in a year, more than 40% of the domestic market. The second is overcapacity: there is more capacity than domestic demand, and the surplus goes to the external market.

The scale of the change within one generation's lifetime: in 1995 the United States produced 24% of world output and China 3%; by 2022 the score was 12% against 32%.

### Lines inside the ribbon

Some countries are better considered not by individual periods but as development stories stretched over decades.

**Korea** offers a ladder in its purest form. 1975: the Pony out of other people's parts. 1986: entry into the US market with the Excel. 1990s: platforms of its own. 2000s: a top-5 world producer. Each stage rested on the previous one: the country moved from assembly using foreign technology to its own platforms and mass exports. The window: political will plus export discipline.

**China** repeated the path in the second generation and turned out to be the only new ladder of the twenty-first century: joint ventures and localization on internal combustion, then electric traction. The window opened twice: first the market (WTO, 2001), then the technology (batteries and software).

**Russia** is taking the exam again. 1970: VAZ under a Fiat licence. 2014–2015: the Lada B/C platform of its own. 2021: the Renaulution plan left AvtoVAZ with a single French platform, CMF-B, and the course towards its own development broke before the sanctions. 2022: the lines stop. 2026: the Volga on Chinese foundations. The full analysis of the chain is in the case below.

**India** runs in parallel: 6.49 million cars in 2025, fourth place in the world. By volume India has already joined the largest producers, but its engineering school has not yet reached a comparable level. At the same time the state's goal is formulated quite concretely: this concerns the PLI programme, to which we will return below. It is the main candidate for the next line of the ribbon.

**T-1. The ribbon of leaders, 1945–2026**

| Years      | Leader                                | What happened                                                                                                                                                                | Cause of the change                |
| ---------- | ------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------- |
| 1945–1970s | United States                         | Post-war monopoly: ~8 of the world's 10.6 million cars (1950); Japan bets on exports and quality, the USSR buys a Fiat licence                                               | demand and incomes                 |
| 1970s      | United States (formally)              | Oil shocks of 1973/1979; Japan becomes the top exporter for the first time (1974); Korea starts with the Pony                                                                | fuel and the structure of demand   |
| 1980–1994  | Japan, later already on American soil | Change of leader: 11.04 million (+14.6%) against a fall of more than 30% in the United States, exports at 54% of output; VER 1.68 → 1.85 → 2.30 million, plants move to the United States | demand and fuel + trade policy     |
| 1994–2005  | United States                         | Return in the year the VER programme closed: the pull of the market and the dollar, not the forge; Japanese plants in the United States; Japan's decline after 1990; NAFTA   | geography of plants + trade policy |
| 2006–2008  | Japan                                 | Crisis of the American market                                                                                                                                                | demand and incomes                 |
| 2009–2021  | China                                 | 13.8 million (2009); WTO and 50/50 joint ventures, growth of the domestic market, electric traction                                                                          | trade policy + support policy      |
| 2022–2026  | China                                 | The world's top exporter: 4.91 million (2023) against Japan's 4.42 million; NEVs 12.9 million                                                                                | technological shift + overcapacity |

*Caption: output shows the geography of plants, not "whose brand"; the ribbon measures all motor vehicles (OICA), while the USSR and Russia rows in the text are passenger cars only (Goskomstat/Autostat). Sources: OICA, Ward's, CAAM, JAMA, JATO.*

![Car output by country, 1960–2025: eleven countries, seven turning points](/upload/upload-why-some-countries-build-cars-fig-1-production-by-country-1960-2025.svg)

Fig. 2 · line chart. Output by country, 1960–2025: eight lines, with the USSR as a separate line up to 1991; turning points are labelled with the causes (1980, 1994, 2009, 2022, 2025). The caption carries the caveats: OICA rows for 1960–1990 for the USSR and China are approximate; the USSR/Russia is counted on passenger cars (Goskomstat/Autostat), the other lines on all motor vehicles (OICA), and the scales are separated by an explicit visual split. Sources: OICA, CAAM, JAMA, KAMA, ASM-Holding.

Sixty years, four changes of first place: Japan in 1980, the United States in 1994, Japan again in 2006, China in 2009. Let us compress them into a single picture.

## The matrix: school and chain

Now let us bring the historical examples together into one scheme. The scheme has two axes: horizontally, the depth of the production chain and the presence of suppliers of one's own; vertically, the engineering school, its own platforms and specialists. Four quadrants, four fates.

**The top right** gathers both forces: the United States, Germany, Japan, Korea and China of the 2020s. This is the top level of development. It is hard to climb here, and often harder still to hold the position.

**The bottom right** is occupied by a chain without a school: Mexico, Brazil, Spain, Slovakia. Here millions of cars are built in plants of world level, but over decades no platforms of their own are created. Spain has been assembling cars since the 1950s, and after joining the Common Market (1986) it became Europe's assembly workshop: output swung from a peak of 3.0 million around the turn of the millennium to 2.0 million by 2012, recovered to 2.9 million and contracted again to 2.3 million by 2025. Slovakia holds the world record for output per head: about 1.1 million cars for 5.4 million people. There is production scale, but an independent engineering school does not arise. That is the ceiling of the assembly model. After 2022, Russia joins this row as well: the plants run on foreign platforms, and in-house development stopped back in 2021 — more on that in the case study below.

**The top left** goes to a school without a chain. Historical insets belong here too: Britain of the 1970s, where the school was preserved and the mass industry given away (the fate of British Leyland), and France, where the school is strong and the scale is leaving: output fell from about 3.5 million in the mid-2000s to 1.4–1.5 million by the mid-2020s.

**The bottom left** is occupied by starting positions: Russia of the 1990s and countries that never began.

No less important on the scheme are the trajectories of countries, not only their position in a given year. Korea 1975→2025 travels from the bottom left corner to the top right in half a century. China 2000→2025 travels the same path twice, first on internal combustion, then on electric traction. Russia 1990→2010→2025 rises towards the middle, and after 2022 drifts right, into the assembly quadrant: the school has stopped developing, and the chain is being rebuilt on foreign platforms and components. A general conclusion follows from these trajectories: a noticeable rise usually happens when a window of opportunity opens. Between windows everyone stands still or slides. Even a strong engineering school cannot grow sharply if the market and the technological conditions do not give it that opportunity. Japan, Korea and China waited for their window for decades. Russia also needs a new window of opportunity, yet for now its industry relies to a large extent on assembling ready-made solutions rather than on development of its own.

**T-2. Eleven countries by seven indicators**

| Country       | Localization | Platforms | Suppliers, tier 2–3 | Personnel | Price/level | Exports | Personnel model     |
| ------------- | ------------ | --------- | ------------------- | --------- | ----------- | ------- | ------------------- |
| United States | ●●●          | ●●●       | ●●●                 | ●●○       | ●○○         | ●●○     | recruits            |
| Germany       | ●●●          | ●●●       | ●●●                 | ●●●       | ●●○         | ●●●     | grows               |
| Japan         | ●●●          | ●●●       | ●●●                 | ●●●       | ●●○         | ●●●     | grows               |
| Korea         | ●●●          | ●●●       | ●●●                 | ●●○       | ●●○         | ●●●     | acquired → grows    |
| China         | ●●●          | ●●●       | ●●●                 | ●●●       | ●●●         | ●●●     | scale + acquisition |
| India         | ●●○          | ●●○       | ●●○                 | ●●○       | ●●○         | ●●○     | grows               |
| Mexico        | ●●●          | ●○○       | ●○○                 | ●○○       | —           | ●●○\*   | assembler           |
| Brazil        | ●●○          | ●○○       | ●○○                 | ●○○       | ●●○         | ●○○     | assembler           |
| Spain         | ●●○          | ●○○       | ●●○                 | ●○○       | ●●○         | ●●○     | assembler           |
| Russia        | ●●○          | ●●○       | ●○○                 | ●●○       | ●○○         | ●○○     | loses               |
| France        | ●●○          | ●●●       | ●●●                 | ●●●       | ●●○         | ●●○     | elite + state       |

*Caption: the author's assessments under the methodology of block 2 (●●● strong, ●●○ medium, ●○○ weak); a colour scale is used in publication. For Mexico the asterisk: exports are large, but these are other people's brands, and the sixth indicator measures the school's exam, not the volume of shipments. France is added beyond the ten as a historical inset: a strong school with a shrinking scale, not the current state of a leader.*

![School and chain: four fates, three trajectories](/upload/upload-why-some-countries-build-cars-fig-3-school-versus-chain.svg)

Fig. 3 · 2×2 matrix. "School × chain": eleven countries and historical insets (Britain of the 1970s), trajectory arrows for Korea 1975→2025, China 2000→2025 and Russia 1990→2025. The author's calculation under the seven indicators; the caption carries the year of the data.

Now comes the Russian example, which shows the difference between assembly and full-fledged production especially clearly.

## The Russia Case: A School Without a Chain

The Russian example allows us to examine the problem in almost laboratory conditions: an engineering school existed, but in a short time the country lost a significant part of its familiar production chain. Let us examine what exactly happened, without gloating and without attempts to excuse obvious weaknesses.

**Ninety-six percent.** Based on 2025 results, AvtoVAZ reported localization of 96% for the Largus, Vesta, Granta, and both Niva models. At first glance, the figure looks impressive. But it refers to a count by nomenclature: a component can be considered Russian if it was assembled in Russia, even when the equipment, technology, or patent belongs to a foreign company. The Ministry of Industry and Trade's point system measures production depth, and by it the three cars diverge like steps: the Granta has 5284 points, the Vesta 4496, and the Moskvich 3 only 2376, against a threshold of about 3700 points, above which the state returns the vehicle recycling fee to the manufacturer. The Moskvich does not even reach the threshold. The difference between roughly 2.4 thousand and 5.3 thousand points vividly shows how far apart simple assembly and deeper production can be.

**Stop-2022.** In the spring of 2022, all assembly lines stopped, including AvtoVAZ's: there were not enough chips and airbags. For the proposed methodology, this is an illustrative case. An assembly line that stops because of a single microchip is an exact portrait of a chain without a tier-2 and tier-3 supplier base. The plants were producing cars, but dependence on individual imported components remained critical.

**The school lived on until the end.** It is important not to oversimplify the picture here: Russia really did have its own platform. The Lada B/C, on which the Vesta is based, was developed by AvtoVAZ engineers with the participation of Renault-Nissan alliance specialists in 2014–2015. But in 2021, a year before the sanctions, the Renaulution plan left the plant only one platform, the French CMF-B. The abandonment of independent development was tied to the alliance's corporate strategy, not to an inevitable course of history. This means the problems began even before 2022.

**Price without level.** Over four years, the Vesta's price in rubles grew by a factor of 1.6–2.2, from 0.9 million at the end of 2021 to 1.46–1.98 million by 2025. In dollars, at recommended prices, the increase is comparable: from about 12 to 19–21 thousand, up 60–70%. Meanwhile, the platform remained a 2015 design. This is a vivid example for the fifth indicator: the price rises while the platform stays the same.

**Rebranding.** The Moskvich 3 replicates the JAC JS4, and the Volga reassembles a Geely under a different name (the brand itself denies the cooperation). This is not the development of the former engineering school, but the use of ready-made foreign models in a situation where in-house solutions are unavailable or will not be ready in time. The reasons must be stated directly: the sanctions shock and the lack of time. Explaining what happened by "financialization" alone would be a stretch. The result is telling: the Moskvich's PST points are half the Granta's, and it does not reach the threshold for recycling fee compensation.

**Age and departure.** The red flag here is not those who did not go to work at the factory: as long as demand for digital skills exceeds demand for metal, engineering graduates will move into IT, and this is the normal functioning of the market, not betrayal. Two signals matter here. The first is the age of personnel: there is no precise public statistics on design bureaus, but the indirect signs are honest. At KAMAZ, the average age of personnel has exceeded 40 (2023), young people make up three out of ten employees, and the plant's design school, according to its chief designer, is sustained by people formed in the 1970s. If the average age of engineers at AvtoVAZ or KAMAZ approaches 55 or higher, it will be a serious sign that the school is not renewing itself and is held up by the older generation. The second signal is the departure of specialists: about 520 thousand STEM graduates per year and 71.5% working in their field look quite decent until you remember that those who left do not appear in these counters at all.

**Chain of causes 2021→2026.** Let us look at this story through the mechanisms already described. 2021 brings a one-time shock: the chip shortage pressures everyone, and Russia has no reserve. 2022 breaks the chain. 2024–2025 raise money and credit: the vehicle recycling fee was raised by 70–85% from October 2024 (Resolution No. 1255), followed by annual indexation, and from December 2025 the methodology was reworked around engine power; for one category the series looks like 300.6 → 556 → 667 thousand rubles over three years. On top of that, a high key rate: car loans become more expensive, the market shrinks. 2026 changes platforms: the revived brands stand on Chinese bases.

**Background.** The new passenger car market for 2025 totals 1.33 million cars, minus 16%. LADA holds 329.9 thousand, minus 24%. Car imports from China fell by a quarter. Both demand and supply are shrinking simultaneously — a combination that is especially hard on the industry.

**Case conclusion.** The problem is not the absence of capable people. Strong mathematical training has been preserved, engineers exist, and the country was developing its own platform as recently as ten years ago. The weak point is the production chain and the demand for independent engineering solutions. There is also a deeper problem — the absence of an agreed goal: import substitution, technological sovereignty, market support via the recycling fee, and the transition to electric propulsion are all declared simultaneously, and these goals pull in different directions. The vehicle recycling fee can protect the market and support prices, but by itself it does not create an engineering school. Without requirements for technology development, such protection risks turning into rent. Instead of investing in engineering teams and development, the country is largely paying for market protection. A "Volga" on a Chinese platform does not square with sovereignty rhetoric. As long as it is not defined for what result the production chain is to be restored, it is unlikely to arise on its own. The opportunity will be seized by whoever understands in advance what goal they are pursuing. Is it reversible? Yes, but only during a period of change: a technology shift, the arrival of suppliers, the opening of a market. Without such a window, the country risks remaining an assembly site, and the buyer paying substantially more for a car.

## China: The Second Miracle and Its Price

After the Russian case, it is useful to look at the opposite example. Over the sixty years under consideration, China is the third and so far the last country that has traveled the entire path from assembling other people's cars to its own engineering school and a developed supplier chain. Moreover, this transition happened twice: first in internal combustion engine vehicles in the 1990s–2000s, then in electric vehicles in the 2010s–2020s.

The scale of China's auto industry is visible across several indicators: 34.5 million vehicles in 2025, up 10% for the year, first place in exports since 2023, 12.89 million EVs produced and 12.87 million sold in 2024, up a third for the year, more than 40% of the domestic market. Behind it stand its own world-class suppliers: CATL makes batteries for half the planet. Ahead lie its own premium brands.

But this rise also has a price, visible in the financial statements. The profitability of China's auto manufacturing industry for 2024 was 4.3%, which is below the country's industrial average: 5.66% for January–August 2026 and 5.9% for 2025. Industry profit for 2024 fell by 8%; 2025 closed at 4.1%, and over the first eight months of 2026 the margin slid to 3.6%, with profit down another 16%. Inside there is a price war driven by BYD, overproduction of capacity, and retaliatory tariffs from the EU and the US. The government propped up demand with a trade-in program for old cars, up to 20 thousand yuan per car through the end of 2024, and in June 2025 Beijing itself intervened in the price war and restricted discounts: competition had spiraled out of control even by Chinese standards.

Both sides matter for our topic. China's leap cost less than is often assumed in the West, but required far greater effort than it may appear from Russia. A low margin does not necessarily mean failure: at a certain stage, Chinese companies may have deliberately sacrificed profit for market share.

![Profitability: scale is not margin](/upload/upload-why-some-countries-build-cars-fig-4-profitability.svg)

Fig. 4 · bars. Profitability: Toyota, VW, BYD, China's industry (3.6% for the first eight months of 2026 versus 4.3% for 2024), AvtoVAZ for a single year alongside. Sources: company annual reports, China's National Bureau of Statistics (NBS); caption with the year of the data.

Seven indicators make it possible to assess production capabilities, but they do not answer another question: for what purpose does a country or company develop them? All seven indicators describe capabilities — what a country knows how to do. But China pursued market share, VW spends about $21 billion on development, and a new car in America costs more than $50 thousand. Capability does not always imply purpose. To discuss purposes, let us add an authorial hypothesis. This is not a measured fact, but a framework for further discussion.

## Goals: environment or flows?

Let us start with countries and turn to corporations later. In all three cases, the rise began not simply with building plants or allocating money, but with a concrete goal that could be translated into practical tasks. Japan of the 1960s–1980s did not formulate "global leadership" or "national revival": such formulations by themselves give engineers no plan of action. The goal was an engineering task that could be measured: fewer defects per thousand cars than the Americans, lower fuel consumption, longer service intervals. And a public consensus: exports were a national project for the whole country, not the business plan of a single Toyota.

Korea moved by milestones, each measurable: the Pony (1975), the Excel on the US market (1986), its own platforms (the 1990s). The rationalization of 1981–1982 under Chun Doo-hwan fixed a segment for each player, and loans went to specific models, not to slogans. China formulated its goal cynically and transparently: technology in exchange for market. The 50/50 joint venture was the legal form of this deal, and every company entering knew the rules. The mercantilist ideology proved effective precisely because it did not hide behind pretty words.

Let us call this the law of purpose: a country gets a chance to rise when its engineering tasks are concrete, openly formulated, and enjoy public support. If, instead of concrete tasks, there remain advertising promises, financial indicators, or several mutually contradictory slogans, a country may miss the right moment: opportunities will appear, but a direction of action will never be chosen.

This hypothesis can be tested on the examples of countries that have already appeared in this article. Mexico did not necessarily "fail": it became a convenient production base for the American market. But the goal of attracting assembly capacity is not the same as the goal of creating one's own engineering school. By the 1970s, the USSR had an engineering school, a production chain, and a significant volume of output. But the goal of "catching up with the West in quantity" ceased to lead the industry forward, and no new goal of comparable clarity appeared. Passenger car output, having soared to 1.3 million by the end of the 1970s (1.33 million in 1980, a peak of 1.35 million in 1987), was stuck at that level until the very end. India today is more interesting than Brazil not in size but in purpose: the PLI scheme pays plants for localizing specific components and producing EVs, over $3 billion over five years, with the task formulated down to the numbers.

Similar logic operates in corporations as well, though it is expressed differently there.

I propose the following hypothesis: there are two models of management. The first is **creating conditions for production**: training personnel, developing suppliers, supporting research and access to markets. Success is judged by the product itself: whether it is getting better and how much its price is growing. The second is **managing financial indicators**: share price, quarterly profit, and return on capital. In both cases, it is worth asking the same question: who invests the money, over what time horizon, and by what criteria does it judge success?

There are data that do not allow everything to be reduced to an opposition between engineering and finance. Volkswagen spends €19–21 billion a year on development: 2025 closed at €19.4 billion after a 7.7% cut, which is 6.7% of the automotive division's revenue, about $21 billion. GM remains the industry leader in R&D share: about 6.3% in 2022, 5.8% in 2023 ($9.9 billion on revenue of $171.8 billion), and for 2025 spending fell to $8.5 billion. Mercedes and BMW each spend $10–13 billion. In other words, major auto groups continue to invest significant funds in engineering development.

But there are also signs that support this hypothesis. The average price of a new car in the US in December 2025 crossed $50 thousand for the first time ($50,614 according to KBB) and remains near the record ($50,562 in September 2026); in China a new car costs about 165 thousand yuan, that is roughly $23 thousand. These markets cannot be compared directly: the mix of cars sold differs. But the difference in average price level is still noticeable. Since late 2023, GM has announced buybacks of its own shares worth $16 billion, and over 2023–2025 it bought back more than $15 billion; on R&D for 2023 it spent $9.9 billion. The result is a contradiction: the company remains a technology manufacturer, but significant sums are also directed to shareholder payouts, not only to product development. Among Japan's big seven, R&D spending holds at 3.92% of revenue, and the May 2026 forecasts promise a nearly twofold drop in operating profit: they will economize, among other things, on the future product. And a historical anchor: the golden age of the European auto industry, the 1950s–1970s, coincided with an era of long, largely public money. I will speak about the same thing, only in cinema, in the conclusion.

Here it is important not to mix two different theses. First: the goals of corporate governance have shifted over two generations from the product to capital; this is the thesis of the framework. The second — "a country pays for other people's products with its own money": a conversation about the trade deficit and reserve currency, a separate story about the US; it hardly applies to Russia and Europe, and here it is not an argument.

If testing this hypothesis on data, it is worth comparing: shareholder payouts (dividends plus buybacks) against R&D and capital expenditures by company and decade; an affordability index, that is, the price of a new car relative to the median salary; the share of engineers on boards of directors, Germany versus the US. As long as comparable data by company and decade do not exist, this framework remains a hypothesis and is not included in the seven-indicator index. An honest caveat: profits and buybacks by themselves do not prove the decline of engineering. This hypothesis makes sense only if it takes both facts into account: yes, VW spends $21 billion on R&D, and yes, a car in America costs more than $50 thousand. The question is not only the size of R&D spending, but who sets the priorities, plans over what time horizon, and rewards management for what.

## Objections Worth Hearing

**"Restyling is also the work of an engineering school."** Yes, if the manufacturer itself refines its own platform. But changing the appearance of a car built on someone else's basis does not yet prove the existence of independent engineering competencies.

**"There are rich countries without an auto industry."** There are. Australia closed its production in 2017 and did not become poor. This test does not measure a country's wealth. It shows what level of industrial complexity the country is able to sustain. A country can live well without an auto industry, but then its ceiling lies somewhere else.

**"But VW spends $21 billion on R&D."** It does, though it is cutting: minus 7.7% in 2025 alone. The objection does not break the framework but strengthens it: the question is not the size of spending, but the horizon: who pays, over what term, and by what criterion of success.

And a general caveat: the methodology is simplified, the T-2 assessments are the author's own, this is not a measuring instrument. This is an analytical model that can be challenged on the merits.

## Conclusion: How to Monitor a Country

I will not attempt to predict what the auto industry will look like in sixty years. Instead, I will offer two tools that will help track changes.

**The first tool is a checklist of seven indicators.** But first ask the zero question: does the country have a concrete engineering goal expressed in understandable tasks and indicators, or do only general promises exist so far? Without an answer to it, the other seven indicators will show the state of the industry but will not explain why it is developing in exactly this way. Then, one question per indicator:

1.  Localization: how much added value remains inside the country?
2.  Platforms: one's own or a license, and from what year?
3.  Suppliers: are there own names in the second and third tiers, or nothing but branches of foreign groups?
4.  Personnel: how many engineers graduate, how many leave, and what is the average age of a design engineer?
5.  Price relative to level: is the car getting more expensive faster than it is improving?
6.  Exports: do the cars go to markets with high requirements or to soft ones?
7.  Personnel model: does the country grow, recruit, absorb, or lose talent?

**The second tool is a question that is not part of the index:** who invests the money, over what term, and by what criteria evaluates the result? Reference books will not answer it, but the outcome largely depends on the answer. If the article is reduced to one thought, it comes out like this: a country makes a leap when it knows how to take advantage of a new window of opportunity; it holds its position thanks to its own production chain, and for its absence the buyer ultimately pays.

Going forward, it is worth watching where new opportunities appear. This may be a change of technology — for example, new generations of batteries and software that is becoming an important part of the car — the opening of markets, or the emergence of tier-2 suppliers. In all three cases examined, the countries began preparing their engineering school even before the opportunity for a leap opened. Today one of the main candidates looks to be India: localization is developing, and the state supports its own competencies in electric vehicles. Opportunities may also arise for the countries of Southeast Asia, as well as for Russia — if sustainable domestic demand for its own developments emerges within the country.

The change in funding sources and planning horizon is noticeable not only in industry. Look at European cinema: the parallel is beyond doubt. There is no collapse—the share of European films in EU cinema attendance holds at about a third; the shift is in concentration: the top of the 2025 box office in Europe was taken by American franchises. More important than the simple similarity here is the coincidence of historical periods: the golden years of Italian, French, and British cinema coincided with the golden era of the European auto industry—both industries lived on long-term, largely public capital. The coincidence of eras by itself proves nothing: both industries could have depended on the common conditions of postwar Europe. But it suggests a direction for analysis: perhaps what changed was not the abilities of the creators, but the sources of financing and the time horizons over which people are willing to invest.

And finally—no longer an argument, but a personal opinion one may disagree with: the pursuit of money can crowd the creative element out of the work. The data in this article allow only a more cautious claim: the sources of financing and the planning horizon may have changed earlier than the abilities of the engineers and designers themselves. But if one day it turns out that these are one and the same thing, I will not be surprised.

This is the author's position, and the proposed methodology is deliberately simplified. Individual estimates and tables can be disputed—that is precisely the point of such an analysis. And the "Moskvich 3" still costs about two million rubles. It is now clearer what goes into that price: the buyer pays not only for the car itself, but also for the weak points of the production chain.