China vs. the Soviet Union: planned vs. reform paths
Two countries began as centrally planned socialist economies. One collapsed. The other became the world’s second-largest economy. How much was the reform strategy, and how much was the starting hand each was dealt?
The starting line
Two countries started in roughly the same place: a planned economy, a one-party state, a promise that the plan would out-produce the market. Half a century later one of those lines is a hockey stick and the other is a cliff. The reflex is to credit the reform strategy, gradual in Beijing, big-bang in Moscow. But before either government chose a strategy, the two starting lines were never the same.
To see why, you need one idea from development economics: the Lewis dual-sector model. An economy with a large, low-productivity agricultural sector holds a reserve of surplus labor, people whose departure from the farm costs almost no lost output. Move them into higher-productivity manufacturing and the whole economy grows fast, almost for free, until the reserve runs dry at the “Lewis turning point.” China in 1978 sat at the start of that transition: four in five people still on the land. The Soviet Union in 1985 had already finished it. Forced collectivization and forced-draft industrialization had emptied the countryside by mid-century. The growth arithmetic follows from the structure: a labor-abundant economy can grow by reallocation, the cheap growth available to almost no policy; a capital-deep, over-industrialized one has already spent that dividend and must grow the hard way, through productivity it does not have.
The historical record on both sides, the Soviet construction from the First Five-Year Plan through the over-industrialized 1980s, and the agrarian China that reform inherited in 1978, is the spine of History Ch.15 (Communist economies). The Lewis tradition itself, from the structuralist founding onward, lives in History of Economic Thought Ch.16 §16.1 (Lewis and the structuralist founding).
Two structures, each on its own terms
“China’s reforms have succeeded in large part because of the structure of the Chinese economy at the start, an overwhelmingly rural, low-income economy with a vast pool of underemployed agricultural labor.”
— after Jeffrey Sachs & Wing Thye Woo, “Understanding China’s Economic Performance”, 1997
The structural reading puts China’s success on its starting hand, and the case is strong. China in 1978 had what Gerschenkron would have called the advantages of backwardness: a surplus-labor reserve to industrialize into, a small state-industrial sector relative to the whole economy, and a planning bureaucracy hollowed out by the Cultural Revolution. Fast, broad, visible gains were sitting on the ground waiting to be picked up. Move peasants off collective farms and let them grow grain for themselves and output jumps without anyone building a single new factory. No policy manufactured that opportunity; the structure handed it over. On this reading the headline lesson, gradualism beats shock therapy, is partly an illusion of attribution: China grew because of where it stood, and the strategy got the credit.
“The Soviet economy was not always a failure. In the 1930s, the 1950s, and the 1960s, planned investment in heavy industry achieved one of the most rapid structural transformations the world has seen.”
— after Robert C. Allen, Farm to Factory, 2003
The “USSR was a basket case” story gets this wrong. Allen’s reconstruction shows Soviet planning was a finished success that had run out of road. Forced-draft industrialization built a superpower out of a peasant empire, won a world war, put the first satellite in orbit, and delivered credible great-power growth into the 1960s. The slowdown was the exhaustion of extensive growth, with no more peasants to move and no more idle land to plow. By 1985 the Soviet Union had already done the transition China was about to begin, which is precisely why the cheap reallocation growth was no longer available. The Soviet structure was a hard inheritance, an economy that had spent its dividend. Whatever reform it attempted would start from a worse position than China’s, before a single decision was made.
What holding strategy aside reveals
Set the strategies aside and the two structures alone predict different feasible paths. A labor-abundant economy at the start of its Lewis transition has a growth dividend to harvest and a place, agriculture, to generate quick, broad wins; a mature, over-industrialized economy has neither. Before either country picks a reform, China’s structure makes a gradual, growth-generating reform feasible and the Soviet structure makes any reform a slog through entrenched heavy industry. The structural reading is right that this matters. But it overshoots if it claims structure decided everything: China still had to choose to release agriculture first, and a clumsier government could have squandered the dividend. Structure set the difficulty; strategy still had to play the hand. China was dealt the easier hand. Whether it also played it better is the next question.
The two countries reached the reform table with opposite philosophies about how fast to change. China would feel for stones; Russia would leap. To judge whether the leap was reckless or forced, we have to put the strongest version of each strategy on the table, starting with the less popular one.
Gradualism vs. shock therapy
“You cannot cross a chasm in two jumps. There was no halfway house between central planning and a market economy. A society can change the basic logic of its economic system only by a comprehensive and rapid program of reform.”
— after Jeffrey Sachs, Poland’s Jump to the Market Economy, 1993
Sympathy runs to gradualism. China won, after all. So start with the architect of shock therapy at full conviction. Sachs is not saying fast is nicer than slow. He is saying a half-reformed economy is a trap, and the only way out is to jump. The question is whether he was wrong, or whether China and Russia were standing on opposite edges of two very different chasms.
Underneath the debate is a single problem. A planned economy has to build, more or less from scratch, the institutions a market runs on: free prices, enforceable property rights, contract law, a financial system. The strategic question is sequencing and speed. Build them in order over years (gradualism), or all at once (big bang). Freeing every price simultaneously in an economy of monopoly suppliers with no competitive entry does not produce instant efficient allocation; it produces inflation and supply collapse, because the price signal only works once the surrounding institutions exist. That is the gradualist’s core worry. The big-bang case accepts it as a cost worth paying to escape something worse.
The reform-era chronology on both sides — Soviet perestroika in 1985–91, the Chinese reform waves from 1978 — lives in History Ch.15 §15.7 (Stagnation and collapse); the Chinese reform at depth, from Deng’s agricultural opening onward, is History Ch.17 §17.1 (Deng’s opening and the agricultural revolution). The older question underneath — can a planned economy compute prices at all, and so can it reform from within? — is the socialist calculation debate, whose lineage from Mises and Hayek through Lange and Lerner sits in History of Economic Thought Ch.6 §6.3 (Mises and the calculation problem).
Each strategy in its own voice
“China did not so much dismantle the plan as grow out of it. The planned economy was held roughly constant in absolute terms while a market economy grew up around it.”
— after Barry Naughton, Growing Out of the Plan, 1995
Naughton’s account is gradualism at its strongest. China never tore down the plan. It froze the planned core — the old output quotas, the old supply relationships, guaranteed — and let a market economy grow up at the margin around it. Agriculture went first, because the household-responsibility system delivered fast, visible gains and built a mass constituency for everything that followed; industry, then prices, then finance came only as the institutions to support each were built. The sequencing was improvised, but it obeyed one discipline: never let the reform produce losers large enough to reverse it. Each step paid for the political room to take the next. That was a strategy for keeping a reform alive long enough to finish.
“A partially reformed economy may be worse than either a fully planned or a fully market one: the half-measure creates rents that a new class of insiders captures and then fights to preserve.”
— after Kevin Murphy, Andrei Shleifer & Robert Vishny, “The Transition to a Market Economy: Pitfalls of Partial Reform”, QJE, 1992
The big-bang argument is the partial-reform trap: a half-reformed economy creates arbitrage rents between its controlled and free segments, and the insiders who capture those rents become the most powerful force against finishing the reform. Gradualism, on this reading, is not a gentle road to the market; it is a stable trap that breeds the oligarchs who will freeze it. Add irreversibility. Do it all at once, before the old nomenklatura can organize to reverse it. And add the constraint Russia actually faced, which China did not: by 1991 the Soviet state and its planning apparatus were disintegrating in real time. There was no stable planned core to “grow out of,” because the plan itself was failing. Naughton’s precondition, a working plan you can freeze and outgrow, was what Russia lacked. Shock therapy was a reformer’s response to a different and worse problem.
“Poland did essentially what the textbook prescribed, and after a sharp initial contraction, output recovered faster and further than in almost any other transition economy.”
— on the Balcerowicz plan; see the GDP map POL panel
One country blocks the lazy inference that “shock therapy = Russian disaster.” Poland — the original Sachs case, the Balcerowicz plan of 1990 — is the canonical shock-therapy success. It freed prices and stabilized fast, took a sharp output drop, and then recovered faster and further than Russia ever did, with its institutions stabilizing. Poland is a data point that sharpens the real variable. The same big-bang strategy produced a stabilized market economy in Warsaw and an asset grab in Moscow. The difference was what each state could enforce while the shock was administered.
Strategy, or recipe?
China’s strategy was good, and the gradualist account is right that dual-track sequencing avoided an output collapse. But the strategy was available to China partly because its structure and its intact state gave it a stable planned core to grow out of. Russia’s shock therapy was a defensible answer to the partial-reform trap and a state that was already dissolving; the gradualist critique that it produced an avoidable output collapse and an oligarchic capture has real force, yet “Russia should have just done what China did” mistakes a strategy for a recipe. The precondition for Chinese gradualism, a capable state and a stable plan to phase out, was precisely what Russia did not have, and Poland shows the same strategy works where that precondition holds. So the headline debate is real, but gradualism versus shock therapy is downstream of something the speed framing never names. To see what, look at what happened to prices and factories on the morning after reform.
Both sides wanted the same destination: a working market that clears. One path got there without a collapse and one triggered one. The difference is in the machinery. Time to look at the steel mill that sold at two prices at once, and the factories Russia handed to its insiders.
The mechanism
Through the 1980s, the same Chinese steel mill sold a ton of steel at one price to fill its plan quota, and at three or four times that price for every ton it made above the quota. Two prices, one factory, one good, at the same moment.
— the dual-track price system
An economist’s first reaction is that this is absurd, an arbitrage waiting to be exploited, a system that cannot hold. It held for a decade. To see why China’s mechanism grew the economy while Russia’s handed it to insiders, you have to understand what running two prices at once does, and what voucher privatization tried to do instead.
Two ideas carry this stage. First, the marginal-price logic behind the dual track: hold the planned quantities and revenues fixed (so no existing claimant loses anything) and let the market track operate only at the margin, where the price clears. Allocation improves where it matters. The next ton is bought and sold at a real price, without the redistribution shock of freeing every price at once. Second, Kornai’s soft-budget constraint, the disease both reforms were trying to cure: a state enterprise that is never allowed to fail has no reason to economize, so the question is always what hardens the budget. China’s township-and-village enterprises and special economic zones grew a hard-budget, market-facing economy outside the state sector. Russia’s voucher privatization tried to harden the budget inside the existing state sector all at once by transferring ownership, but ownership passed to insiders who kept extracting rents.
Let the plan fix quota $\bar{q}$ at planned price $p_0$, and let any output above $\bar q$ trade at the market price $p_m$. Each existing claimant is held at least as well off as before reform (the plan track is guaranteed), so no one has a veto-worthy loss; meanwhile every marginal unit is produced and allocated at $p_m$, where $p_m$ equates marginal cost and marginal value. The reform is therefore Pareto-improving and incentive-compatible, with efficiency gains at the margin and no losers at the core. That is why the dual track was politically survivable when a single-track price jump was not.
Guarantee everyone what they already had at the old price, then let them sell anything extra at the real price. Nobody is made worse off, so nobody fights to stop it, and because the extra ton trades at a price that means something, the economy starts allocating resources properly at the edge. Russia tried to fix the whole thing at once by handing out ownership papers. But if the state can’t make the new owners face real consequences, the papers just relabel the same soft, rent-seeking factories under private names.
The records on both sides — the dual-track, TVE, and SEZ takeoff against the Soviet voucher program — sit in History Ch.15 §15.8 (What planning could and couldn’t do) and, for the Chinese mechanism at depth, History Ch.17 §17.2 (TVEs, SEZs, and the first industrial takeoff). Kornai’s soft-budget diagnostic is where the Austrian calculation tradition ends up, the reason it concluded a plan cannot economize. That lineage is in History of Economic Thought Ch.6 §6.4 (Hayek and the knowledge problem).
Two mechanisms, each defended
“The dual-track approach achieves efficiency-enhancing reform without creating losers. It is a mechanism for reform without losers, and that is why it was politically feasible where a single-track liberalization was not.”
— after Lawrence Lau, Yingyi Qian & Gérard Roland, “Reform without Losers”, JPE, 2000
Lau, Qian, and Roland make the Chinese mechanism’s strongest case. The dual track is a theoretically clean transition path: by guaranteeing pre-reform claims on the plan track and liberalizing only the margin on the market track, it delivers allocative efficiency and leaves no existing claimant worse off, which is why it survived politically. TVEs and SEZs are the institutional embodiment of the same logic: hard-budget, market-disciplined growth that never had to defeat the state-sector incumbents first, because it grew up beside them. China built an economy where the argument about ownership did not have to be won before growth could start.
“The central goal of privatization was to depoliticize economic life, to get assets out of the hands of politicians and bureaucrats, fast and irreversibly, even at the cost of an imperfect initial allocation.”
— after Maxim Boycko, Andrei Shleifer & Robert Vishny, Privatizing Russia, 1995
And the voucher program was not naive. Given a collapsing state with no capacity to run a multi-year sequenced transition, the priority was to get assets out of state hands fast and irreversibly: depoliticize the economy before the nomenklatura could recapture it, create a property-owning class with a stake in the market’s survival, and accept an imperfect initial allocation because any private allocation hardens the budget constraint more than continued state ownership. The partial-reform trap made the patient gradualist alternative look like a stable corruption equilibrium. The reformers knew the first allocation would be ugly; they bet that ugly-and-private beat tidy-and-state in a country where the state was vanishing. That the bet failed where the state could not enforce the new property rights is the verdict’s business, but it was a reasoned bet under brutal constraints.
The cleanest isolation in the comparison
Hold the goal constant — harden the budget constraint, get prices to clear — and the mechanisms separate cleanly. The dual track grew a hard-constraint economy at the margin without a redistribution shock; voucher privatization tried to transform the existing sector all at once and, absent a state able to enforce the new property rights, handed control to insiders who kept the soft budget alive under private names. The reason the dual track worked is that China had a functioning state and a stable plan to phase out at the margin; the reason vouchers produced capture is that Russia had neither a stable plan to grow out of nor a state capable of enforcing the property rights it was creating. Kornai’s diagnostic was right about the disease on both sides, soft budgets everywhere, but the cure depended on a state strong enough to administer it. The mechanism difference is real, and it is downstream of the state-capacity difference.
Structure, strategy and mechanism each keep pointing at the state itself. China kept its state and reformed its economy. The Soviet Union tried to reform both at once. That single difference in sequencing may explain more than all the economics combined.
The politics
“Cross the river by feeling for the stones. We must keep a cool head… and on the question of the political system, we will not copy the West.”
— a composite of Deng Xiaoping’s reform-era sayings, as commonly quoted (the ‘feeling for the stones’ phrase is usually traced to Chen Yun)
“Perestroika without glasnost is impossible. We need democratization of the whole of society, this is the decisive condition for the success of restructuring.”
— after Mikhail Gorbachev, Perestroika, 1987
The difference sits in two sentences from two leaders. Deng reformed the economy and refused to touch the party; Gorbachev opened the politics and the economy at the same time. Same starting system, opposite sequencing. China kept its state and the reform worked. The Soviet Union loosened its state, and the reform and the state dissolved together. The last question is how much of the divergence is this single choice, and whether it was a choice at all.
One idea closes the argument: state capacity. Markets need a state that can enforce contracts and property rights and keep order; a reform that destroys state capacity destroys the precondition for the markets it is building. North’s work on credible commitment makes the point sharp. A market needs to believe that the rules will still hold tomorrow, which requires a state both willing and able to bind itself. Under the Acemoglu-Robinson extractive-versus-inclusive lens, the operative question is not which institutions were more inclusive but whether the reform preserved enough state capacity to make any new institutions binding at all. A property right no one can enforce is not a property right.
The Soviet political-economy chronology — glasnost and perestroika together, then the 1991 dissolution — is History Ch.15 §15.7 (The era of stagnation and the collapse). The credible-commitment and institutions lineage, from North through Acemoglu, lives in History of Economic Thought Ch.15 §15.4 (New Institutional Economics: Williamson, North).
Two sequences, each at its strongest
“China escaped the poverty trap through directed improvisation, central direction that empowered, rather than replaced, decentralized local experimentation.”
— after Yuen Yuen Ang, How China Escaped the Poverty Trap, 2016
Ang gives China’s sequencing its strongest reading. Holding the party intact preserved a state that could enforce the new market rules, keep order through a wrenching transition, and credibly promise that the regime which started the reform would still be standing to honor it. “Directed improvisation” let local governments try TVEs and SEZs and scale whatever worked, under central political control that kept the experiments from spinning into chaos. The political continuity was load-bearing for the economic success. The state that stayed standing is the reason the agrarian dividend could be harvested, the dual track administered, and the new property rights enforced.
“The economic and the political blockages in the Soviet system were the same people. To move the economy, Gorbachev judged he had to break the bureaucracy’s grip, and that meant opening the politics.”
— the case for glasnost-and-perestroika-together (after Gorbachev, Perestroika, 1987)
And the Soviet case at its strongest is that Gorbachev’s simultaneity may not have been a free strategic blunder. Soviet economic stagnation was entangled with political ossification: the planning bureaucracy and the military-industrial complex were political actors who blocked economic reform from inside. Gorbachev may have judged, defensibly, that economic restructuring required political opening to break the bureaucratic veto, that glasnost was the lever to mobilize public pressure against the rent-capturing nomenklatura the partial-reform-trap literature warned about. The gamble failed. The political opening accelerated the state’s dissolution faster than the economic reform could deliver gains. That does not make it foolish in advance. It makes it a high-variance bet in a system where the economic and the political blockages were the same men.
The variable underneath all the others
China reformed the economy while holding the state and the party capable; the agrarian dividend (Stage 1) could be harvested, the dual-track strategy (Stage 2) administered, the mechanism (Stage 3) enforced, all because the state stayed standing. The Soviet Union’s simultaneous political and economic liberalization, whether a free choice or a forced response to a system whose political and economic blockages were identical, removed the state capacity that every other variable depended on, and the reform’s preconditions dissolved with the state. The variable the comparison isolates is political sequencing and state capacity. “Gradualism versus shock therapy,” the speed frame, is the wrong axis. The right one is “did the reform preserve a state capable of enforcing the transition?” China’s gradualism worked because it kept the state; Russia’s shock therapy struggled partly because the state was already gone. A planned economy can be reformed gradually if the state stays capable and the structure offers early wins; where those conditions fail, shock therapy is a defensible response to a worse problem.
The finding that political continuity was load-bearing for China’s economic reform is one short step from “authoritarianism plans better,” and that step is not warranted here. The autocracies we observe are the successful ones, out of a far larger pile that failed; the developmental democracies of Korea and Japan ran their own state-led miracles; and whether autocracy generally plans better is a separate, genuinely contested question, taken up at Are autocracies better at long-run planning?. This comparison measures one thing, growth-and-output reform success, and the claim is narrow: a reform needs a capable state to enforce it. The economic verdict implies no political one.
Where this leaves us
We started with two lines: a hockey stick and a cliff, from two countries that began as the same kind of economy. The reflex credits the reform strategy — gradual in Beijing, big-bang in Moscow — and declares gradualism the winner. Structure set the difficulty: China had a surplus-labor dividend and a hollowed bureaucracy; the Soviet Union had spent its dividend and entrenched its planners. Strategy still had to play the hand, and China’s fit its structure, but the gradualist precondition, a stable plan to grow out of and a state to administer it, was exactly what Russia lacked, and Poland proves the same strategy can succeed where that precondition holds. The mechanism that avoided collapse, the dual track that ran two prices at once, was downstream of state capacity. And underneath all of it sat the political choice: keep the state, or risk it.
The answer is a conditional. “Gradualism beats shock therapy” is real, but it depends on initial conditions and, above all, on state capacity. The right axis is sequencing, and whether the reform kept a state capable of enforcing the transition. Hold the Soviet side at its own strongest throughout — Allen’s real industrial achievement, the partial-reform trap, a Gorbachev facing a genuine bind — and the lesson is that the same medicine needs a body strong enough to survive it.
- Initial conditions set the difficulty, China’s surplus-labor dividend versus the Soviet Union’s exhausted, over-industrialized structure.
- Strategy still had to play the hand. China’s gradual sequencing fit its structure; Russia’s big bang answered a partial-reform trap and a collapsing state.
- Mechanism diverged. The dual track grew a hard-budget economy at the margin; voucher privatization handed the old soft-budget firms to insiders, and the difference was downstream of state capacity.
- State capacity was the variable underneath all the others. The verdict is conditional: a plan can be reformed gradually if the state stays capable; where it cannot, shock therapy is a hard answer to a harder problem.
This comparison is one deep, two-case instance of a larger question, why are some countries rich and others poor?, whose institutions strand is the general form of the state-capacity finding here. Whether China’s reform model can outlive its own success, and whether autocracies plan better in the long run, are separate questions left to their own walkthroughs.