Corporate Global Catastrophic Risks (C-GCRs)

byHaukeHillebrandt 2mo30th Jun 201913 comments


Epistemic status

Very uncertain and speculative. I don’t excessively hedge my claims throughout for clarity’s sake (‘better wrong than vague’).


Are corporations causing global catastrophic risks (C-GCRs)?

Here, I argue for this to be true, based on the following claims:

  1. Corporations grow ever larger over time.
  2. Corporations grow exponentially more powerful over time.
  3. Corporations cause exponentially more regulatory capture over time.
  4. Corporations cause exponentially larger externalities over time.

If all these claims are true and cannot be falsified, then it follows that the externalities of big, powerful, unregulated corporations will increase over time to the level of GCRs.

I then argue for the following corollaries:

  1. Corporations are already the distal cause and main driver of anthropogenic and emergent technological GCRs such as global catastrophic biological risks (GCBRs).
  2. C-GCRs are a bigger threat than GCRs from state actors.
  3. C-GCR reduction is more effective than targeted GCR reduction because it is broader, large in scale, neglected, and solvable.

I suggest concrete policy proposals to reduce C-GCRs, by diversifying corporate ownership, enforcing corporate taxes, and optimizing funding for regulatory agencies.

1. Are corporations growing ever larger?

Economic theory suggests that corporations try to maximize profits. Incentives create strong selective pressure for corporations to grow (e.g. because they benefit from economies of scale).[1]

Empirically, corporations have become much bigger especially in the last ~20 years, because of cross-border mergers & acquisitions:[2]

  • Revenues of the Fortune 500 increased from 58% to 73% of US GDP
  • Bigger, Fortune 100, corps grew even faster: from 33% to 46% of GDP. Their share of Fortune 500 revenues increased from 57% to 63%.[3] Some of it due to increased revenues from abroad.
  • Of corporations with $1bn+ revenues the top 10% now capture 80% of profits—up 1.6x.[4]
  • Industries become “increasingly dominated by ‘superstar firms’ with high profits and a low share of labor in firm value-added and sales”.[5] For instance, Walmart has >2 million employees‚ FAANG corps only on the order of 100k.
  • Markups, the ratio of the price to the marginal cost of production, have also increased.

This trend should continue: some corporations’ market capitalizations are ~$1tn.[6] This is a proxy for market valuation—the discounted sum of all future profits. Profits and revenues correlate, so the market predicts that corporate revenues and thus power will increase. Multinationals can grow bigger than non-multinationals because of cross-border M&A.[7]

As a rule, corporations grow larger over time.

2. Do corporations optimize to become exponentially more powerful over time?

Power is roughly proportional to how many resources an agent has to achieve its goals.

Recently, researchers have proposed that states’ “budgets” (tax revenues) and corporations’ “budgets” (revenues minus their profits) are comparable in that they are crude, but useful proxies for their power (despite corporations and states spending on different things).[8]

For instance, Shell’s revenue/budget to sell fossil fuels is $272bn. About half of Boeing’s $96bn revenue comes from its defense contractor subsidiary—the amount of power to sell military equipment.

Big state actors (e.g. U.S., EU, China) have the biggest budgets (+$1tn).[9] But, surprisingly, some corporations now have ~$0.5tn budgets. Walmart has higher revenues than Spain or Australia and Apple has higher revenues than Belgium or Mexico. 71 out of the 100 most powerful actors are corporations with combined budgets of $10tn vs. $18tn in combined budgets of the 29 states (see Appendix; spreadsheet).

If a corporation spends just 0.1% of its $0.5tn annual budget on things that might cause negative externalities, this equates to $0.5bn. For instance, Apple spends >$1bn/year on marketing. Lobbying budgets could increase to a similar level.

Also, corporations sometimes combine their power. For instance, if lobbying for certain favorable regulations benefits a whole industry, this creates a free-rider problem. Corporations routinely solve this problem by sharing lobbying costs and throwing budgets together to fund industry bodies and trade associations. This way, corporations become even more powerful. In contrast, state power is curtailed by mandatory and entitlement spending taking up an increasing share of their budgets (e.g. >70% in the US).[10]

Overall, corporate power has increased to the level of many state actors.

3. Do corporations cause exponentially larger externalities over time?

As corporations grow, they create exponentially larger:

  • externalities, both positive and negative
  • consumer surplus, but also harm to consumers

Historically, big corporations robustly increased net social welfare—outweighing negative externalities.[11],[12],[13] Positive externalities: Larger corporations contribute disproportionately to the economic performance of countries - they are more productive, pay higher wages, enjoy higher profits and are more successful in international markets.[14] Uber’s easily quantifiable consumer surplus is ~$3bn.[15] Even the attention economy, vilified for wasting trillions in people’s time, might actually create billions in consumer surplus through more content and more efficient ads.[16]

But Big Business (such as Big Oil, Tobacco, Alcohol, Pharma, Food, Meat, Agro, Tech, Media, Finance), also creates increasingly negative externalities and harm consumers.

Empirically, we have seen Big Business increasingly playing a large part in the following negative externalities:

  • The Global Obesity Epidemic: 2 billion people are overweight[17] (though some costs are internalised by consumers)
  • The Global Tobacco Epidemic: 1 billion people might die prematurely[18] (though some costs are internalised by consumers)
  • Factory farms: 50 billion animals/year killed.[19]
  • 2008 financial crisis: >100% GDP cost for the U.S. ($50k-$120k for every U.S. household) and caused 500k excess cancer-related deaths worldwide (through increased unemployment).[20],[21]
  • The prison-industrial complex incarcerates many people and lobbies[22] (but see [23])
  • Addictive technologies (e.g. social media, smartphones) affect billions and have been called an emerging public health problem.[24],[25]

In theory, corporate profit-maximization functions should correlate strongly with the social welfare maximization function. But if consumer choice does not correlate perfectly with (long-term average, total) utility (of all morally-relevant agents), then even a small divergence might result in side-effects causing large negative externalities. Because destroying value is easier than creating value, negative externalities can become very high as evidenced by the examples above. Consider climate change: this is not merely a coincidence, but the result of something inevitably causing harm as a by-product of ruthless optimization.

Thus, as a rule, as corporations grow they cause exponentially larger positive externalities, consumer surplus and utility, but also cause exponentially larger negative externalities, harm consumers and disutility.

4. Do corporations cause regulatory capture proportional to their size?

In economics, governments are to reduce businesses’ negative externalities via regulation. But governments are increasingly unable to regulate large corporations. Why?

  1. Historically, corporate and state power was coupled, because corporate tax on profits increased government budgets proportionally. But average global corporate tax rates are decreasing, from 47% (GDP-weighted: 39%) in 1980 to only 23% (GDP-weighted: 26%) in 2018.[26] Corporations actively try to avoid taxes.[27]
  2. Governments’ capacity to regulate has decreased: between 1979 and 2005, both US House committee and Government Accountability Office staff decreased by ~40%; Research Service staff, which provides nonpartisan policy analysis to lawmakers, by 20%.[28]
  3. Corporations can use lobbying for regulatory capture[29]. The economics literature has long worried about increasing regulatory capture.[30] Generally, lobbying has increased in recent years[31]. In the US, corporations spend about $3bn/year lobbying the government[32]. In the EU, lobbying has also increased in recent years.[33] In particular, Big Tech’s and Finance’s lobbying budgets have increased.[34],[35]

In general, increasingly powerful corporations can promote their interests through lobbying and evade increasingly weak government regulation through regulatory capture.

5. Are anthropogenic and emergent technological GCRs a form of C-GCR?

Because humanity has survived for >200k years, natural GCRs (e.g. asteroids, pandemics) only have a 0.1%-1% risk of causing extinction per century. This might be 10x less than man-made GCRs (e.g. nuclear war, engineered bioweapons, AI, climate change).[36]

Corporations spend more on R&D than governments.[37] Generally, corporations are relevant actors in areas of anthropogenic GCRs from emerging technology such as risks from AI[38], GCBRs[39], climate change, and war.

  • Nuclear weapons: The U.S. plans to spend $1.7 trillion on nuclear weapons in the coming years[40]–a share of which will go to corporate contractors.
  • Bio-risk: Corporations have received billions in recent years to work on synthetic biology, a field changing the bio-economy, worth several trillions.[41],[42] Since 2001, US biodefense research spending has increased 20-fold, even though it does not (yet) meet the definition of a biomedical military–industrial complex.[43]
  • AI-risk: Corporations are the main driver of developments in AI and spend billions on R&D.[44] The level of technical AI expertise at leading AI labs is substantially higher than that within the US military.[45] Competitive pressure to develop AI is the only reason for AI risk because it takes away the option of slowing down AI development until we have a good solution to the alignment problem.[46]

Corporations also lobby on GCR-relevant issues:

  • Climate change: World Energy expenditure is in the trillions. [47],[48] Thus, it makes sense that fossil fuel companies have spent >$1Bn on misleading climate-related lobbying.[49] Climate change might be a GCR edge case: its direct effects might not be an x-risk, but its indirect effects could cascade to GCR-level given that there is a 10% risk of >6°C warming (see [50] for discussion).
  • Nuclear War / Great power war: Corporations are lobbying to get NATO countries to spend 2%/GDP on military.[51] The US ICBM program was partially created due to corporate lobbying.[52] Defense company Lockheed Martin has the second most lobbyists in DC (31 in-house lobbyists + 53 connections with lobbying firms).[53] In France, India, UK, and the US, corporations make atomic bombs.[54] Corporation made +$2bn in profit in the last 10 years, and there are many safety risks, while the National Nuclear Security Administration is understaffed.[55] China’s and Russia’s nuclear weapons programs are government-owned and controlled, but some talk about a Russian nuclear-industrial complex.[56]
  • Biorisk: Biotech-pharma has diluted the Biological Weapons Convention verification protocol and influenced the US to reject it.[57]
  • Other emerging tech: Fukushima has been cited as a case study for emergent technological GCRs: the Japanese nuclear sector had revolving doors to the regulator, “destined to result in [...] regulatory capture.” [58],[59]

Thus, corporations deserve special attention when horizon-scanning for emerging technological GCRs as they are their distal cause.

6. C-GCR reduction is more effective than targeted GCR from emerging tech reduction

Reducing C-GCRs is a more generalized intervention than targeted reduction of GCRs from emerging tech. Its broad cross-cutting nature makes it more effective.[60]

Because corporations will create as of yet unknown emerging tech that causes GCRs, C-GCR reduction is good for horizon-scanning and preventative GCR reduction. Regulators or altruists who focus on particular GCRs play Whac-A-Mole with corporations. Every industry can create its own GCR - for instance, see research on GCRs from chemicals.[61] Even if one is particularly concerned about just one GCR from say AI, C-GCR prevention is more effective than targeted GCR reduction, because the distal source of it comes from corporations.

What can be done? I go through some policies in the next sections.

7. Policy solutions to decrease corporate GCRs

1. Breaking up big corporations

Because corporations cause exponentially more disutility as they grow, breaking up big corporations (or preventing M&A) to lower their power would reduce the size of their externalities. But this is intractable, as countries will worry too much about their economic competitiveness. There is no consensus amongst economists on whether governments should regulate the size of big corporations more generally (e.g. through preventing M&A or breaking them up).[62],[63]

2. Increasing corporate taxes for bigger corporations

Corporate taxes have some advantages, but they are not usually considered to be Pigovian[64] (and might also have other drawbacks[65]). But in light of C-GCRs, corporate taxes become Pigovian because corporations are usually causing large externalities. Increasing (the progressiveness of) corporate taxes (i.e. the bigger the corporation the higher the tax)[66] might reduce C-GCRs.

But currently corporate tax is regressive. Why? Because bigger corporations are better at avoiding tax by shifting profits overseas, and thus have a lower tax rate than smaller firms.[67] A politically tractable solution might be to reduce corporate tax avoidance.

A solution might be to tax the self-assessed value of a corporation’s domestic legal entity. [68] At the self-assessed value would ba strike price that they’d be required to sell it to anyone. For instance, Apple pays little corporate tax in the UK, because Apple UK does not make a lot of profit. But getting Apple to give a true estimate of the value of Apple UK Ltd and putting say a 5% tax might be harder to avoid.

3. Mission hedging to fund regulators

A foundation whose mission is to stop climate change can invest their endowment using ‘mission hedging’ by investing in fossil fuel stocks.[69] This way it has more money to give to organizations that combat climate change when more fossil fuels are burned, fossil fuel stocks go up and climate change will get particularly bad. When fewer fossil fuels are burnt and fossil fuels stocks go down, the foundation will have less money, but it does not need the money as much. Generally, increasing investment in objectionable corporations creates a hedge around an actor's mission, which sometimes, maximizes expected utility.[70]

Similarly, governments can use mission hedging to reduce C-GCRs. For instance, governments can set up a sovereign wealth fund that invests in an index fund that covers all corporations weighted by market capitalization or revenue. Dividends from individual corporate stocks could then be used to fund regulators so their budgets increase with corporate size. In other words, if Apple revenue/power increases and Apple stocks generate higher returns, this would increase the staff/budget of a department within a regulatory agency tasked with regulating Apple. Similarly, the sovereign wealth fund could also own stocks of foreign firms to fund intelligence agencies to monitor foreign corporations.

4. Diversifying corporate ownership through sovereign wealth funds and reducing home bias

If the ownership of corporations would be more diversified more, then there would be fewer incentives to create externalities. In other words, if everyone were a shareholder of Big Oil, then shareholder value might be maximized by reducing emissions. This is because both profits and negative externalities are shared more widely and nobody has an advantage from profiting from untaxed externalities.

This might also reduce arms races: if a technology’s up- and downsides are shared, then it decreases adversarialism and increases collaboration.[71] Relatively few individuals sometimes own and control major AI corps (FAANG).[72],[73] In contrast, ~25% of Baidu is owned by 10 Western funds (Vanguard, Blackrock, etc).[74]

Governments setting up sovereign wealth funds that invest in an index fund that covers all corporations weighted by market capitalization or revenue would diversify ownership.

This policy might be politically tractable. Countries such as China now allow fully foreign-owned enterprises,[75] increasing government-ownership of corporations has recently been suggested,[76] and sovereign wealth funds might increase tax receipts[77] and create a hedge against risks to the domestic economy. Note that governments being minority shareholders across an industry would be different from state-owned enterprises that often create more externalities than private firms.[78]

Another way to diversify ownership of corporations might be to reduce home bias, where countries and people hold suboptimal amounts of foreign equity. US investors have 78% of their equity portfolio in U.S. stocks, which are only a third of world market portfolio, by capitalization.[79] Reducing home bias as an intervention might be tractable because getting home biased investors to diversify ownership of corporations might also benefit them financially. This might be an alternative to windfall clauses.[80] Governments could grant tax incentives to their citizens invest in foreign equities or give incentives for diversifying their portfolio.

Coda: Are corporations optimization demons?

2.5bn years ago stromatolites changed the atmosphere from a CO2-rich to O2-rich through photosynthesis, because they had no competition.[81] A corporation might not be a superintelligence,[82] but it might create one. We should not let corporations lead on AI governance as has been suggested.[83] Usually, CSR/PR is relatively small relative to lobbying activities and externalities corporations create (e.g. fossil fuel companies will sometimes come out for a carbon tax, but then spend money on anti-climate lobbying; corporations will invest a small amount of money into AI safety but spend much more AI R&D).

In sum, I think it might be useful to think of corporations as dangerous optimization demons which will cause GCRs if left unchecked by altruism and philanthropy.[84]

Appendix: Top 100 state and corporate actor by revenue ≈ power

Adapted from Barbic et al. "States versus Corporations"

Spreadsheet here

Banking M&A chart:


[1] Breaking down the barriers to firm growth in Europe

[2] States versus Corporations

[3] FiveThirtyEight - Big Business Is Getting Bigger

[4] McKinsey Global Institute ‘Superstars’—The dynamics of firms, sectors, and cities leading the global economy

[5] David Autor - The Fall of the Labor Share and the Rise of Superstar Firms

[6] List of public corporations by market capitalization

[7] List of multinational corporations - Wikipedia

[8] States versus Corporations

[9] Wikipedia — List of public corporations by market capitalization

[10] Wikipedia - US federal budget #Mandatory spending and entitlements

[11] Steven Pinker - Enlightenment Now

[12] Tyler Cowen - Big Business

[13] Big Is Beautiful

[14] Breaking down the barriers to firm growth in Europe

[15] Using Big Data to Estimate Consumer Surplus

[16] The Economics of Attention Markets

[17] Health Effects of Overweight and Obesity in 195 Countries over 25 Years

[18] The Global Tobacco Epidemic

[19] Factory farming - 80,000 Hours

[20] Economic downturns, universal health coverage, and cancer mortality in high-income and middle-income countries, 1990–2010

[21] Financial crisis may have caused 500,000 cancer deaths worldwide: study

[22] Prison–industrial complex - Wikipedia

[23] Are private prisons driving mass incarceration?

[24] Problematic smartphone use: Digital approaches to an emerging public health problem

[25] Adam Alter - Irresistible

[26] Tax Foundation - Corporate Tax Rates Around the World

[27] Corporate Tax Avoidance Remains Rampant

[28] Lindsey - The Captured Economy

[29] Stigler - The Theory of Economic Regulation

[30] Regulatory Capture - A Review

[31] Modern Lobbying: A Relationship Market

[32] Tyler Cowen - Big Business

[33] LobbyFacts Database

[34] Google, Amazon, and Facebook all spent record amounts last year lobbying the US government

[35] Igan - Bank Lobbying: Regulatory Capture and Beyond

[36] Toby Ord — Will We Cause Our Own Extinction? Natural versus Anthropogenic Extinction Risks

[37] Global private and public R&D funding

[38] The Malicious Use of Artificial Intelligence: Forecasting, Prevention, and Mitigation

[39] Schoch-Spana et al. - Global Catastrophic Biological Risks: Toward a Working Definition

[40] Congress Of The United States Congressional Budget Office—Approaches for Managing the Costs of U.S. Nuclear Forces, 2017 to 2046

[41] Center for Biosafety Research and Strategy, Tianjin University, China — Synthetic biology: Recent progress, biosafety and biosecurity concerns, and possible solutions

[42] Synthetic Biology in the Driving Seat of the Bioeconomy

[43] A biomedical military–industrial complex?

[44] Artificial intelligence: The next digital frontier? - McKinsey

[45] - US AI policy profile

[46] Thomas Sittler - A shift in arguments for AI risk

[47] World energy expenditures

[48] Energy expenditure, economic growth, and the minimum EROI of society

[49] - Big Oil's Real Agenda on Climate Change

[50] John Halstead - Is climate change an existential risk?

[51] Defense Contractors Say Russian Threat Is Great

[52] Who's Really Driving Nuclear-Weapons Production?

[53] Tyler Cowen - Big Business

[54] Don't Bank on the Bomb

[55] Light penalties and lax oversight encourage weak safety culture at nuclear weapons labs

[56] The Russian Nuclear-Industrial-Complex

[57] Unwarranted Influence? The Impact of the Biotech-Pharmaceutical Industry on U.S. Policy on the BWC Verification Protocol

[58] Classifying global catastrophic risks

[59] A Nuclear Complex? A Network Visualization of Japan’s Nuclear Industry and Regulatory Elite

[60] Nick Beckstead - A Proposed Adjustment to the Astronomical Waste Argument

[61] Global Catastrophic Risks from Chemical Contamination

[62] IGM Forum - Breaking Up Large Tech Companies

[63] IGM Forum - Breaking up European Champions

[64] Why Tax Corporations?

[65] Good recent review paper on corporate taxes effects

[66] Why isn't the corporate income tax progressive?

[67] Corporate Tax Avoidance Remains Rampant

[68] Glen Weyl – Radical Markets

[69] Brigitte Roth Tran - Divest, Disregard, or Double Down?

[70] Hauke Hillebrandt - A generalized strategy of 'mission hedging'

[71] Amanda Askell on the 80k podcast

[72] The Top 5 Google Shareholders

[73] The Top 6 Shareholders of Facebook

[74] Baidu Shareholders

[75] China unveils draft law to allow fully foreign-owned enterprises

[76] Mariana Mazzucato - The Entrepreneurial State

[77] Paul Christiano on Sovereign Wealth Funds

[78] When Governments Regulate Governments

[79] Literature review on Home Bias

[80] Allen Defoe - AI Strategy, Policy, and Governance

[81] Stromatolites

[82] Corporations vs. superintelligences - Arbital

[83] Jade Leung - Why Companies Should be Leading on AI Governance

[84] Paul Christiano - More realistic tales of doom