Resource dividends and social wealth funds: lessons from Alaska, Norway, Iran and Mongolia

If a society owns a stream of income that does not depend on anyone's labour — oil, gas, minerals or, in some future proposals, a share of the returns to automation — it has to decide what to do with it. It can spend it on public services, save it in a fund for future generations, or pay it directly to citizens as a dividend. Real cases of all three exist, and they are among the best natural experiments available for thinking about "social wealth funds" and "AI dividends". This study compares four: Alaska's Permanent Fund Dividend, paid annually since 1982; Norway's Government Pension Fund Global, which pays no dividend at all; Iran's universal cash transfer, introduced in December 2010 to compensate for the removal of energy and bread subsidies; and Mongolia's resource-financed payments to children and then to all citizens, which were scaled back after a few years. The cases show that direct dividends can be popular, administrable and compatible with continued work, but that their durability depends less on the idea of sharing resource wealth than on the rules that link payments to revenue — and on whether those rules survive politics.

Key findings

• Alaska pays everyone, but the amount is political. Since the first $1,000 dividend in 1982, Alaska has paid an equal annual dividend to every eligible resident. The amount has ranged from $331.29 (1984) to $3,284 (2022); in 2025 it was $1,000, paid to 618,863 applicants at a total cost of about $620 million.
• The Alaska dividend has not reduced employment. Jones and Marinescu (2022), using a synthetic control method, find no effect on employment and a 1.8 percentage point (17%) increase in part-time work, consistent with the cash stimulating the local economy.
• Since 2018 the Alaska fund pays out by rule, but the dividend is set by appropriation. Legislation enacted in 2018 (SB 26) fixed the annual draw from the fund's earnings as a percentage of its average market value — 5.25%, later 5% — and the dividend now competes with the rest of the state budget for that draw.
• Norway saves and spends through the budget instead. The Government Pension Fund Global, created by law in 1990 and first funded in 1996, is invested abroad; a fiscal rule limits government spending of fund income to the expected real return, estimated at about 3% a year. The fund passed 20,000 billion kroner in 2024 and owns almost 1.5% of all shares in the world's listed companies.
• Iran's transfer was universal and large, then eroded. In 2011, its first full year, Iran's transfer amounted to 6.5% of GDP and about 29% of median household income. Salehi-Isfahani and Mostafavi-Dehzooei find no evidence that it reduced labour supply, except among young people with weak ties to the labour market. But payments were set higher than the subsidy savings could finance, the deficit was partly covered by printing money, and the nominal amount was frozen while prices rose nearly 28-fold between 2010 and 2025.
• Mongolia's scheme was driven by election promises. Resource-financed payments to children began in 2006, were extended to all citizens in 2010 and reverted to children only in 2012. Payments were based on election pledges rather than actual mineral revenue, at times exceeded that revenue and were financed by borrowing; public support for direct disbursement fell from a peak of 25% to under 10%.
• The design lesson is about rules, not slogans. Dividend schemes survive when payouts are tied to a transparent, rule-based draw from a savings fund; they fail when payments are promised before revenues exist or are left to annual bargaining.

Research question and scope

This study asks three questions. How have the main real-world mechanisms for distributing resource income — dividends, sovereign wealth funds with fiscal rules, and universal transfers replacing subsidies — worked in practice? What does the evidence show about their effects on work and welfare? And what design choices determine whether such schemes are durable, fair and fiscally sound?

The scope is four national or sub-national cases chosen to span the main design options. We draw on official data (Alaska's Permanent Fund Dividend Division, Norges Bank Investment Management, Alaska's legislature), on peer-reviewed studies (Jones and Marinescu on Alaska; Salehi-Isfahani and Mostafavi-Dehzooei on Iran) and on policy research (Yeung and Howes on Mongolia; Moss, Lambert and Majerowicz on "oil-to-cash"). We do not cover the many sovereign wealth funds that neither pay dividends nor operate under explicit fiscal rules, and we do not estimate what an "AI dividend" would be worth.

Background

The resource curse and the case for direct distribution

Resource-rich countries often struggle to convert resource wealth into broad prosperity. Revenue that flows to governments rather than citizens can weaken accountability, invite corruption and encourage volatile spending. One proposed remedy, promoted by the Center for Global Development under the name "oil-to-cash", is to pay a share of resource revenue directly to citizens and tax part of it back, so that citizens experience resource wealth as their own and governments must ask them for money. Yeung and Howes summarise the argument: by putting resource rents in the hands of the people, incentives to undermine public institutions would be removed, corruption avoided, and accountability and transparency would grow. Whether this works in practice is precisely what the cases below test.

From resource dividends to social wealth funds

The same logic underlies proposals for "social wealth funds" or "AI dividends": if a growing share of national income accrues to owners of capital — including automated systems — a public fund holding part of that capital could pay every citizen a dividend. Resource dividends are the closest real-world analogue, because they distribute income from an asset that is collectively owned and does not depend on recipients' work. The analogy has limits: resource revenues are finite and volatile, and the institutions that hold them were built for that purpose.

Method and sources

For each case we document the legal basis, the rule (or absence of one) linking payments to revenue, the size of payments, the evidence on effects and the political history. Figures for Alaska come from the Permanent Fund Dividend Division's official payment summary; for Norway from Norges Bank Investment Management; for Iran from the peer-reviewed study and working paper by Salehi-Isfahani and Mostafavi-Dehzooei and a 2026 analysis by Salehi-Isfahani; and for Mongolia from Yeung and Howes. Where we could not verify a figure from a primary or peer-reviewed source, we describe the point qualitatively.

Findings: Alaska's Permanent Fund Dividend

Origins

Alaska voters approved a constitutional amendment establishing the Permanent Fund in 1976. The first dividend legislation, enacted in 1980, would have given each adult resident $50 for every year of residency since statehood in 1959. That scheme was challenged as unconstitutional, and in Zobel v. Williams (1982) the US Supreme Court struck it down. The legislature responded by authorising equal dividends for all residents meeting a six-month residency requirement; the first dividend of $1,000 was distributed on 14 June 1982. The residency rule has since been changed several times — in 1989 the legislature extended it to 24 months, a court reduced it to 12 months in 1990, and in 1993 the residency period was changed to a calendar year.

Payments over time

Source: Alaska Department of Revenue, Permanent Fund Dividend Division, Summary of Dividend Applications & Payments. Amounts are nominal.

Over 1982–2025 the division paid 24,919,293 dividends totalling about $31.3 billion, and the cumulative dividend for someone eligible in every year was $52,831.41. The table shows the central weakness of the design from a recipient's point of view: the dividend is unpredictable. It halved between 2015 and 2016, tripled between 2021 and 2022, and fell by almost 60% between 2024 and 2025.

The 2018 rule

For most of its history the dividend was calculated by a statutory formula from the fund's earnings. As oil revenue declined, the state began to rely on fund earnings for general spending. Legislation enacted on 8 May 2018 (SB 26) required the Permanent Fund Corporation to determine each year an amount available for appropriation equal to 5.25% of the average market value of the fund over the first five of the preceding six fiscal years, later reduced to 5%. This "percent of market value" draw protects the fund's principal and smooths payouts, but it means that the dividend and public services are now financed from the same pool, with the split decided by the legislature. The dividend's size has become one of the most contested questions in Alaska politics.

Effects on work and welfare

Because the dividend is universal and permanent, it offers rare evidence on how an unconditional cash payment affects labour markets at scale. Jones and Marinescu (2022) compare Alaska with a synthetic control built from other states using the Current Population Survey. They find that the dividend had no effect on employment and increased part-time work by 1.8 percentage points, or 17%. Non-tradable sectors, which depend on local demand, showed a more positive employment response than tradable sectors. The authors interpret this as a general-equilibrium effect: cash stimulates local spending, which raises labour demand and offsets any reduction in individual labour supply. They conclude that a universal and permanent cash transfer does not significantly decrease aggregate employment.

The limits are important. The dividend has typically been a modest share of annual income, and it is paid once a year. A payment large enough to live on, paid monthly, could have different effects.

Findings: Norway's Government Pension Fund Global

Saving instead of distributing

Norway took the opposite approach. According to Norges Bank Investment Management (NBIM), it was decided early that revenue from oil and gas should be used cautiously to avoid imbalances in the economy. In 1990 Parliament passed legislation creating what is now the Government Pension Fund Global, and the first money was deposited in 1996. The fund invests only abroad. It holds small stakes in around 7,200 companies across most countries and industries — almost 1.5% of all shares in the world's listed companies — making it one of the world's largest funds. Its value passed 10,000 billion kroner in 2019, 15,000 billion in 2023 and 20,000 billion in 2024.

The fiscal rule

Government budget deficits are covered with transfers from the fund, and surpluses are added to it. Politicians have agreed on a fiscal rule under which, on average, the government spends only the equivalent of the expected real return on the fund, which NBIM describes as estimated at around 3% a year. Oil revenue is thus phased into the economy gradually, and in principle the fund's capital is never spent. Norway's citizens receive no cheque, but the fund's returns finance a substantial share of public spending.

What Norway shows

Norway shows that collective ownership of resource wealth need not mean individual distribution. Its model concentrates decisions in Parliament and the central bank, relies on high trust in institutions and on a political consensus around the fiscal rule, and treats the fund as an intergenerational asset rather than a personal entitlement. The trade-off is that citizens' stake is indirect: the fund's benefits arrive as public services and lower taxes, which are harder to see and easier to dispute than a dividend.

Findings: Iran's 2010 cash transfer reform

Design

Iran's parliament passed the legislation for a subsidy reform in January 2010, but the law was not implemented until December of that year. The government then raised prices of bread and energy products by factors ranging from two to nine, removing subsidies estimated at $50–60 billion, about 15% of GDP, and simultaneously released cash it had deposited in dedicated household bank accounts. The initial plan was to compensate only the poorest third of households, but because identifying them proved administratively impractical, the government paid everyone. The amount was 455,000 rials per person per month.

Scale

In 2011, the first full year of the programme, transfers amounted to 6.5% of GDP and about 29% of median household income — making Iran's programme, for a time, one of the largest universal cash transfers in the world. Transfers equalled 4.9% of per capita expenditure for people in the top quintile and 49.3% for those in the bottom quintile. The working paper notes that the programme contributed significantly to lowering poverty and inequality.

Effects on work

Senior politicians criticised the programme as handouts that would "foster beggars", and press reports claimed that poor workers and small farmers were abandoning work. Salehi-Isfahani and Mostafavi-Dehzooei tested this using household panel data, exploiting variation in the intensity of treatment (transfers relative to prior household spending) and in the timing of registration, since about 30% of the population had to wait three months for the registration system to reopen. With the exception of young people, who have weak ties to the labour market, they find no evidence that the transfers reduced labour supply. Service-sector workers appear to have increased their hours, perhaps because some used the transfers to expand their businesses.

Financing and erosion

The weakness was financial. In Salehi-Isfahani's 2026 account, the amount paid proved too high to be financed by the savings from the reform, and the resulting deficit was partly financed by printing money. Inflation, which had been about 10% a year before the energy price rise, accelerated several-fold in 2011 and 2012; the working paper records a rise from 12.4% in 2010 to 34.7% in 2013, with other causes including tightening international sanctions from July 2011. Public support for further energy-price increases — which would have preserved the transfers' real value — evaporated. Energy prices were frozen, and so was the nominal value of the transfer. By 2025 the consumer price index had increased nearly 28-fold relative to 2010, reducing the original transfer to a small fraction of its purchasing power.

From 2025, as war loomed and then broke out in June, new funds were injected into the programme, this time without corresponding energy-price increases. The programme became targeted: it now excludes the top income decile and pays 3,000,000 rials a month in deciles five to eight and 4,000,000 rials in the bottom four deciles, alongside electronic food vouchers.

Findings: Mongolia's resources-to-cash experiment

Design and history

Mongolia, a resource-rich young democracy, is perhaps the only developing country that has actually introduced a resources-to-cash scheme. In 2006 it introduced resource-financed payments for all children; in 2010 the scheme was expanded, with larger payments to be provided to all citizens; in 2012 it reverted to children only. Mongolia retained universal child transfers but announced that it would delink them from resource revenue, and the new sovereign wealth fund it was establishing would not be earmarked, even partially, for cash transfers.

What went wrong

Yeung and Howes find that the transfers significantly reduced poverty and inequality, but that the scheme was poorly designed and implemented. The link to resource revenue was weak in practice: payments were based on election promises rather than actual resource earnings. Child payments were introduced during the 2004 election campaign; in 2008 parties competed over the size of the transfers they would authorise; by 2011 the election law was amended to stop this practice. At some points cash transfers exceeded mineral revenue, and the gap was met by borrowing, increasing debt and possibly adding to inflation. Public support fell: in opinion polls by the Sant Maral Foundation between 2008 and 2014, "direct disbursement" as the preferred use of mining wealth declined from a peak of 25% to under 10%. Interviewees described the experiment as wasteful and irresponsible, and even as an abuse of the political process.

The authors' conclusion

Yeung and Howes call Mongolia a cautionary tale, not a refutation. One poorly designed instance should not be used to dismiss resources-to-cash, but the case shows that the very ease of handing money out also makes it easier for this option to blow the budget. They are also sceptical of the claim that resource dividends strengthen accountability through taxation: in neither Alaska nor Mongolia are the transfers taxed, nor did the tax regimes change when payments were introduced.

Comparing the four cases

Design trade-offs

Dividend, fund or budget?

A dividend makes citizens' ownership visible and hard to divert. A fund with a fiscal rule preserves wealth for future generations and smooths spending. Budget spending allows resource income to finance public goods that individual dividends cannot buy. Alaska combines a fund with a dividend; Norway combines a fund with budget spending; Iran and Mongolia paid out without a buffer fund. The two cases with a buffer fund have lasted.

Rule or discretion?

Payments tied to a transparent rule — a fixed percentage of a fund's market value, or of a real return — are predictable and harder to capture. Payments left to annual political decision invite bidding wars, as in Mongolia's elections and in Alaska's annual budget fights. A rule does not end politics: Alaska's legislature still decides how much of the rule-based draw becomes dividends.

Volatility and predictability

Resource revenues are volatile. A dividend that follows revenue closely will be unpredictable, as Alaska's payment history shows. A dividend that is smoothed requires a large fund and a credible rule. For recipients, predictability may matter as much as the amount: the Kenyan basic income study suggests that expected duration changes behaviour.

Financing and inflation

Universal transfers financed by money creation rather than real revenue can trigger inflation that erodes their value and their political support. Iran's experience shows how quickly a popular transfer can lose its purchasing power when its financing is unsound and its nominal value is frozen.

Universal or targeted?

Iran initially intended to target the poorest third, abandoned the plan as impractical, and paid everyone; fifteen years later it moved back towards targeting. Universal payments are administratively simple and politically robust but expensive; targeted ones concentrate resources but require data that governments often lack.

Individual cash or collective goods?

Dividends suit societies where trust in government spending is low and administrative capacity for direct payments is high. Funds with fiscal rules suit societies with strong institutions and consensus about intergenerational saving. There is no evidence that one model is superior everywhere.

What the evidence does not show

• It does not show that resource dividends cure the resource curse. The accountability mechanism proposed by oil-to-cash advocates — taxing dividends back to make citizens scrutinise government — has not operated in Alaska or Mongolia, where dividends are untaxed.
• It does not show what a full basic income from a fund would do. Alaska's dividend is a fraction of living costs. The labour-market evidence cannot be extrapolated to payments large enough to live on.
• It does not show that Norway's model would work elsewhere. Norway's fiscal rule rests on institutional trust and political consensus that are not easily transplanted.
• It does not provide a clean estimate of Iran's inflationary effect. The subsidy removal, the transfer financing and international sanctions all coincided; disentangling them is contested.
• It does not tell us how an AI dividend would behave. Returns to automation would accrue to private owners unless a state acquired a stake, and their size, volatility and taxability are unknown.

Implications for institution design

1. Create the fund before the payment. Durable dividend schemes (Alaska) and durable savings schemes (Norway) both rest on an accumulated fund. Promising payments from expected future revenue, as Mongolia did, invites borrowing and collapse.
2. Fix the draw by rule and publish it. A transparent, rule-based draw — a percentage of average market value or of expected real return — protects capital and limits bidding wars. The rule should be hard to change, for example by requiring supermajorities.
3. Decide explicitly how the draw is split. If both a dividend and public services depend on the same draw, as in Alaska since 2018, the split should be governed by a stable formula, not renegotiated each year.
4. Index or adjust payments. Iran's frozen nominal transfer lost almost all its value. Payments meant to be a lasting share of national wealth need a mechanism to keep their real value or an honest decision to let them shrink.
5. Separate electoral cycles from payment levels. Mongolia's decision to bar campaign competition over transfer sizes came too late. Independent bodies or statutory formulas can insulate payments from election promises.
6. Treat proposed "AI dividends" as fund-design questions. Before debating the amount of an AI dividend, designers must answer how a public fund would acquire stakes in automated capital, what draw rule it would follow, and how it would handle volatile returns.

Open questions

• How should Alaska balance the dividend against public services within the percent-of-market-value draw, and can any formula command lasting consensus?
• Would Norway's model survive a prolonged period of low or negative real returns?
• Can a targeted, digitally delivered transfer in Iran regain the credibility lost when the universal transfer eroded?
• Could a social wealth fund financed by taxes on capital or automation replicate the political robustness of resource funds, which rest on clearly identifiable natural assets?
• What is the minimum institutional capacity needed for a resource dividend to improve accountability rather than undermine it?

Sources

1. Alaska Department of Revenue, Permanent Fund Dividend Division. "Summary of Dividend Applications & Payments." pfd.alaska.gov
2. Alaska Department of Revenue, Permanent Fund Dividend Division. "Historical Timeline." pfd.alaska.gov
3. Alaska State Legislature (2018). SB 26: An Act relating to the earnings of the Alaska permanent fund and the earnings reserve account. akleg.gov
4. Supreme Court of the United States (1982). Zobel v. Williams, 457 U.S. 55. justia.com
5. Jones, D. and Marinescu, I. (2022). "The Labor Market Impacts of Universal and Permanent Cash Transfers: Evidence from the Alaska Permanent Fund." American Economic Journal: Economic Policy 14(2): 315–340. doi:10.1257/pol.20190299
6. Norges Bank Investment Management. "About the fund." nbim.no
7. Salehi-Isfahani, D. and Mostafavi-Dehzooei, M. H. (2018). "Cash transfers and labor supply: Evidence from a large-scale program in Iran." Journal of Development Economics 135: 349–367. doi:10.1016/j.jdeveco.2018.08.005
8. Salehi-Isfahani, D. and Mostafavi-Dehzooei, M. H. (2017). Cash Transfers and Labor Supply: Evidence from a Large-Scale Program in Iran. ERF Working Paper 1090. erf.org.eg
9. Salehi-Isfahani, D. (2026). "Iran Cash Transfers in a Time of War and Inflation." Tyranny of Numbers, 5 August 2026. djavadsalehi.com
10. Yeung, Y. and Howes, S. (2015). Resources-to-Cash: A Cautionary Tale from Mongolia. Development Policy Centre, Australian National University. doi:10.2139/ssrn.2661202
11. Yeung, Y. and Howes, S. (2015). "Resources to cash: a cautionary tale from Mongolia." Devpolicy Blog, 22 October 2015. devpolicy.org
12. Moss, T., Lambert, C. and Majerowicz, S. (2015). Oil-to-Cash: Fighting the Resource Curse through Cash Transfers. Center for Global Development. cgdev.org
13. Egger, D., Haushofer, J., Miguel, E., Niehaus, P. and Walker, M. (2022). "General Equilibrium Effects of Cash Transfers: Experimental Evidence From Kenya." Econometrica 90(6): 2603–2643. doi:10.3982/ECTA17945
14. Banerjee, A., Faye, M., Krueger, A., Niehaus, P. and Suri, T. (2023). Universal Basic Income: Short-Term Results from a Long-Term Experiment in Kenya. Working paper. PDF