Soft Budget Constraint: When a Safety Net Exists, Accountability Distorts

The conceptual tool for this discussion is called the “soft budget constraint.” It can help you further understand public affairs.

However, it also applies to private affairs. You’ve certainly heard of situations like this: a family whose child is already an adult indulges in high consumption, accumulating massive debts from credit cards, Huabei, Jiebei, and the like, but the parents always bail her out.

The first time, the parents said the child was still young, lacked social experience, and would improve later.

The second time, the parents again felt sorry for her, and said, “Let’s pay it off for her; after all, we have money.”

The third time, the parents were already struggling, but they couldn’t just watch their child’s credit record get ruined.

The fourth time, the child proactively reminded her parents that it was time to help her repay the debt.

This is the same principle as the over-budget projects you see in companies. Originally, a project had a budget of ten million, but by the halfway point, eight million had already been spent. The person in charge said, “We’ve already spent so much; wouldn’t stopping now be an even greater waste?” So the headquarters approved another five million. Three months later, it still wasn’t working, and they said, “We’ve come this far; it’s the final push, we can’t fail at the last minute.” So another five million was approved…

Eventually, the money ran out, and the project remained unfinished. But their report was well-written: ‘paths were explored, experience was accumulated, and the team was trained.’

Behind these two incidents is the same mechanism: the one who decides to spend the money is not the one who ultimately pays the bill.

Soft Budget Constraint: Concept and Essence #

The “soft budget constraint” was first introduced by Hungarian economist János Kornai in the late 1970s and early 1980s [1]. He was then studying state-owned enterprises in Hungary and Yugoslavia, and discovered a particularly widespread anomaly: some enterprises were extremely inefficient, losing money year after year, and by market rules, should have gone bankrupt a hundred times over, yet they simply refused to die.

Why? Because the state wouldn’t let them die. No matter how much you lost, the authorities would find a way to keep you alive: granting tax exemptions, providing low-interest loans, or, if all else failed, directly issuing subsidies… Anyway, someone would fill the gap.

A budget is, by its nature, a hard constraint. You can only do as much as the money you have allows; if you run out of funds, you’re out. Long-term losses for an enterprise indicate a flawed business model; you are not creating value, but consuming it. Such enterprises should cease to exist, making way for those that can create value. This is the law of nature.

But if you believe someone will rescue you every time you incur a loss, then a budget is merely a soft constraint for you.

Kornai discovered that the higher the subjective probability that external forces will fill your deficits, the softer your budget constraint becomes. Put simply, this means that as long as you believe “someone will always take care of things if trouble truly arises,” every decision you make today will be bolder in borrowing, more adventurous, and less concerned with costs.

In a 2003 review paper [2], Kornai and his collaborators further proposed that soft budget constraints are not exclusive to socialist and transitional economies, but also apply to market economies. At its core, it’s a “dynamic commitment problem”: when an enterprise first submits its budget, it already foresees that if problems arise, the government (or its superiors) will provide assistance. Thus, the commitment between them is merely temporary; future budgets will inevitably be increased.

In other words, even if you vow beforehand, “I will never bail you out,” when the crunch comes, you often find that you still have to rescue them, thereby breaking your word.

Financial Crises and “Too Big to Fail” #

The most famous example is the 2008 financial crisis, which introduced the world to the phrase “too big to fail.”

At the time, Lehman Brothers collapsed first, and market panic spread rapidly. Insurance giant AIG was also on the verge of collapse. Had it fallen, countless financial institutions worldwide would have been implicated. Consequently, the Federal Reserve and Congress successively injected massive capital into AIG.

In hindsight, the government’s move might have been unavoidable, as it would otherwise have led to a catastrophe. But on the other hand, is it reasonable to use taxpayers’ money to bail out a private financial institution? Is this still capitalism? When you made profits, you didn’t share your bonuses with us; how is it that when you’re losing money and on the brink of collapse, you ask us for help? So, when small businesses fail, it’s called survival of the fittest, but when large institutions get into trouble, it’s called systemic risk?

Regulators learned two lessons from 2008.

First, some institutions genuinely cannot be allowed to suddenly fail, because they would drag the entire system down.

Second, but bailouts cannot be too comfortable. If you protect shareholders, management, and creditors all at once, that’s tantamount to telling all large companies: “Go ahead and bet boldly; if you win, it’s yours; if you lose, it’s the system’s problem.”

When Silicon Valley Bank collapsed in 2023, the regulators’ actions were much more sophisticated [3].

Under conventional rules, U.S. federal deposit insurance only covers up to $250,000 per account, with any amount exceeding that being the depositor’s loss. However, Silicon Valley Bank’s impact was simply too great: if depositors weren’t protected, bank runs could easily spread to other small and medium-sized banks, and a large number of tech companies might be unable to pay salaries… Panic would breed greater panic.

Therefore, relevant authorities invoked the “systemic risk exception,” announcing protection for all deposits, including large deposits exceeding $250,000.

But note, the government only protected depositors, not the bank itself. Silicon Valley Bank’s shareholders lost everything, executives were replaced, and some unsecured creditors were not protected. I’m saving the system, not rewarding your recklessness…

So, will the banking industry learn its lesson and make more cautious decisions in the future? Not necessarily. As long as the market believes “if it truly escalates to systemic risk, the government will still intervene,” this expectation will factor into ex-ante decisions.

This mechanism is dual, operating across both time and space, and related theories have won two Nobel Prizes.

Temporally, Norwegian economist Finn E. Kydland and American economist Edward C. Prescott proposed [4] that the government should indeed commit ex-ante to “not rescuing anyone who gets into trouble,” so that everyone acts cautiously. However, once an event truly occurs, from the ex-post perspective, a bailout is often more rational than allowing a collapse. Thus, the government changes its mind.

And so, the market already foresees that the government will change its mind. Consequently, the most rational ex-post rescue precisely encourages the boldest ex-ante risk-taking. Kydland and Prescott were awarded the Nobel Memorial Prize in Economic Sciences in 2004.

Spatially, French economists Emmanuel Farhi and Jean Tirole discovered [5] that as a financial institution, your smartest move might not be to reduce your own risk, but rather to make your risk similar to everyone else’s — because as long as everyone is crowded onto the same boat, making the same mistake, for example, all engaging in high-risk “maturity mismatch” operations, then if that boat starts to sink, regulators, to prevent the entire economy from going down with it, will have no choice but to intervene.

This is the financial version of “the law doesn’t punish the masses.” One person running a red light will be fined, but if ten thousand people run a red light together, the red light has to surrender. Tirole later received the Nobel Memorial Prize in Economic Sciences in 2014.

Your loss is your failure; but if you can transform your loss into a systemic problem, then it’s no longer your failure, but a skill to make others pay for you.

Power and Scale: The Allure of Budget Maximization #

Even more alarmingly, you can proactively leverage this ability. Not only do larger entities not collapse, but “bigness” itself is a reward.

American financial economist Michael C. Jensen, in his corporate governance research, proposed the “empire building” mechanism [6], pointing out that managers are incentivized to expand the size of their organizations. The larger the scale, the greater the budget, personnel, assets, and influence under their control.

Note that, this pursuit often runs counter to the interests of shareholders. Shareholders are concerned with profit, while managers are more concerned with “how large a domain I oversee.” This is why some companies, even without good projects, still engage in mergers and acquisitions, expansion, hiring more people, and building taller headquarters.

If this is true for professional managers in a market economy, what about within the government? What if the money is purely “public” property, with no shareholders at all? You would then head towards what we discussed in the previous lecture: the “Huang Zongxi’s Law.”

This phenomenon occurs across all times and cultures.

American economist William A. Niskanen proposed a model called the “budget-maximizing bureaucrat” [7] — where bureaucrats pursue not profit, but the scale of the budget they manage: the more money they control, the larger their domain, the greater their power and the brighter their prospects.

To be fair, this mechanism can sometimes be beneficial, providing positive externalities. Especially for developing countries during periods of economic growth, officials proactively expanding budgets, initiating large projects, and improving infrastructure constitutes economic mobilization. Activities like building roads, constructing bridges, and developing industrial parks, if left solely to spontaneous market evolution, might not progress quickly enough. Some economists believe that it was precisely the healthy competition among local governments centered on growth and investment promotion that fostered China’s rapid economic growth.

Economists Hongbin Li and Li-An Zhou refer to this as a “promotion tournament” [8]: local officials compete like athletes on GDP growth rates, and whoever has the most impressive figures is more likely to be promoted. Their research found that the probability of promotion for provincial officials is significantly positively correlated with local economic growth rates.

This is actually a form of progress; promotion based on performance is better than promotion based on connections. However, the budget constraint here is very soft. Government-led investment has become systematized, with “local government financing vehicles” established in various regions, and “urban investment bonds” issued. Officials can easily mobilize funds from banks for any large project.

During periods of rapid growth, almost anything yields positive externalities, but investment-driven growth can only take you so far. At a certain stage, investment exhibits sharply diminishing marginal returns. Yet, officials continue to aggressively pursue performance, and whether those projects will be profitable or loss-making in the long run, or leave behind a mess, will eventually come to reckoning.

You’ll find that past achievements have transformed into debt and excess capacity.

So, with declining investment efficiency, substantial debt, and excess capacity, should spending become more cautious in the future? Not at all. Empirical research by Chinese economists has uncovered a vicious cycle here: bank credit support depresses enterprises’ capacity utilization rates, yet, enterprises with poorer operational efficiency are more likely to receive more credit, and state-owned enterprises, in particular, find it harder to exit [9].

Japan provides an earlier precedent [10]. Between 1991 and 2001, during Japan’s “Lost Decade” following the asset bubble collapse, when the economy was mired in stagnation and deflation, banks, to avoid exposing bad loans, instead prolonged the lives of enterprises that should have gone bankrupt. Consequently, these “zombie firms” that refused to die hogged credit and suppressed prices, dragging healthy enterprises into the grave with them.

This is the insidious power of soft budget constraints. Markets are supposed to reward efficiency, but now inefficiency has ironically become a reason “to lend a hand.” Projects that “shouldn’t have been invested in” in the first place have now become “projects that cannot be stopped.”

Small failures are educated by the market; large failures educate the market.

Soft Budget Constraints: The Institutionalization of Moral Hazard #

There’s an older concept in economics called “moral hazard” [11], originally jargon in the insurance industry, meaning that when an individual doesn’t bear the full consequences of their actions, they become bolder and less cautious. Soft budget constraints are essentially the organizational version of moral hazard.

This structure is far more prevalent than you might imagine.

For instance, why are US college tuition fees becoming increasingly expensive? We previously discussed “Baumol’s Cost Disease” as one reason — but Baumol’s Cost Disease cannot fully explain the exorbitant extent of tuition increases, where a year’s college tuition can exceed a typical family’s annual income. In fact, this is also related to soft budget constraints.

As early as 1987, U.S. Secretary of Education William Bennett proposed that the more generous government student loans became, the more aggressively college tuition would rise, because students, backed by federal loans, would no longer be as price-sensitive. This assertion became known as the “Bennett Hypothesis.” A 2019 paper quantitatively proved this hypothesis: for every $1 increase in federally subsidized loans, college tuition would rise by approximately 60 cents [12].

Healthcare also suffers from this problem. Health insurance is essentially also a form of soft budget constraint: when medical expenses are paid by insurance companies and the government, neither doctors nor patients are as concerned about the price of medications. The trend of modernization is for goods to become cheaper and labor more expensive, yet the prices of medical equipment and drugs in the U.S. have not gotten cheaper.

Economist Milton Friedman addressed this issue long ago [13], coining a widely circulated concept called the “Four Ways to Spend Money.” He argued that spending your own money on yourself prioritizes both savings and effectiveness; whereas spending someone else’s money on someone else’s behalf is neither a concern for cost nor effectiveness, leading to the greatest waste.

Actually, you don’t need an economist; ordinary people also understand this principle. It’s just that ordinary people often don’t know whose money is being spent, or whose affairs are being managed.

Leveraging Soft Budget Constraints: Mastering Two Narrative Abilities #

To fully exploit soft budget constraints, and freely spend other people’s money, you need to master two narrative abilities.

The first is “making a fuss” (闹), which means framing your own failure as everyone’s problem —

“If you don’t give me an additional budget, the system will shut down.”

“If you let me collapse now, all previous investments will be wasted.”

“If you don’t save me, suppliers will demand payment, employees will lose jobs, customers will complain, and leadership will face difficulties explaining.”

What truly matters here is not logic, but volume; the bigger the fuss, the better. Some things are “allocated by making a fuss.” The crying child gets the milk, not because they are hungrier, but because they made everyone in the room hear their hunger.

The second is “packaging” (包装), which means presenting your own interests as essential for the security of the broader situation —

“We are not a loss-making enterprise; we are a pillar industry affecting the livelihoods of tens of thousands of people.”

“This isn’t overcapacity; it’s a strategic reserve for supply chain security.”

“We are not asking for subsidies; this is about protecting employment, protecting livelihoods, and protecting local stability.”

As long as you can frame “I’m losing money” as “if you don’t save me, there will be trouble”; and “I want more” as “you need me to continue existing,” your inefficiency becomes leverage, your failure becomes a credential, and consequently, the basis for your next resource application.

These narratives are not necessarily entirely false. But once the situation reaches this point, those who honestly live within their means and bear their own problems have all become marginalized individuals without systemic importance.

Beware of Grand Narratives: Who Pays the Bill? #

As we discussed earlier, grand narratives are most afraid of accountability. A simple question, “Who pays?”, can help you see many things clearly.

“Stability,” “employment,” “livelihoods,” “supply chain security,” “systemic risk”… These are indeed important, but none of them have clear beneficiaries or responsible parties. Whenever someone discusses these grand terms, you should be wary: they are likely preparing to spend other people’s money.

“The child is still young,” “the elderly have it tough,” “we’re all family,” “I can’t bear to”… These are indeed genuine emotions. But whenever someone appeals to your emotions, you should understand that more often than not, you’ll be footing the bill for this round.

Love must have boundaries; assistance must come with conditions; support is not a blank check; tolerance for error is not exemption from responsibility; power and responsibility should be commensurate. Soft budget constraints are man-made systems, and other people’s money is also finite. The laws of nature ultimately demand that we respect hard constraints.

【Concluding Poem】

A budget is inherently a ruler, But human sentiment bends that rule. Save an inch today, And tomorrow debts become mountains. To rescue, the price must be clear; Only through failure is hardship truly known. Consequences belong to their owner, And compassion should have boundaries.

注释

[1] Kornai, János. “The Soft Budget Constraint.” Kyklos 39, no. 1 (1986): 3–30.

[2] Kornai, János, Eric Maskin, and Gérard Roland. “Understanding the Soft Budget Constraint.” Journal of Economic Literature 41, no. 4 (2003): 1095–1136.

[3] Federal Deposit Insurance Corporation, “FDIC Acts to Protect All Depositors of the former Silicon Valley Bank,” Press Release PR-19-2023, March 13, 2023; 及 U.S. Government Accountability Office, GAO-25-107023, 2025.

[4] Kydland, Finn E., and Edward C. Prescott. “Rules Rather than Discretion: The Inconsistency of Optimal Plans.” Journal of Political Economy 85, no. 3 (1977): 473–492.

[5] Farhi, Emmanuel, and Jean Tirole. “Collective Moral Hazard, Maturity Mismatch, and Systemic Bailouts.” American Economic Review 102, no. 1 (2012): 60–93.

[6] Jensen, Michael C. “Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers.” American Economic Review 76, no. 2 (1986): 323–329.

[7] Niskanen, William A. Bureaucracy and Representative Government. Chicago: Aldine-Atherton, 1971.

[8] Li, Hongbin, and Li-An Zhou. “Political Turnover and Economic Performance: The Incentive Role of Personnel Control in China.” Journal of Public Economics 89, no. 9–10 (2005): 1743–1762. 另见周黎安《中国地方官员的晋升锦标赛模式研究》，《经济研究》2007 年第 7 期。

[9] Ma, Hongqi, Xinxiang Mei, and Yuan Tian. “The Impacts and Potential Mechanisms of Credit Support with Regard to Overcapacity.” Resources Policy 68 (2020): 101704. 另见江飞涛、耿强等关于地区竞争、补贴与产能过剩的实证研究（《中国工业经济》）。

[10] Caballero, Ricardo J., Takeo Hoshi, and Anil K. Kashyap. “Zombie Lending and Depressed Restructuring in Japan.” American Economic Review 98, no. 5 (December 2008): 1943–1977.

[11] Pauly, Mark V. “The Economics of Moral Hazard: Comment”，American Economic Review 58, no. 3 (1968): 531–537.

[12] Bennett, William J. “Our Greedy Colleges.” New York Times, Feb 18, 1987. 实证见 Lucca, David O., Taylor Nadauld, and Karen Shen. “Credit Supply and the Rise in College Tuition.” Review of Financial Studies 32, no. 2 (2019): 423–466.

[13] Friedman, Milton, and Rose Friedman. Free to Choose. 1980.