Permanent Trust Fund Investment and Payout Rules

What the evidence says

The best-covered outcome is whether resource-revenue sovereign wealth funds change government fiscal behavior, and the answer here is not settled. Multiple studies of natural-resource funds associate fund establishment with a smoothing effect on government spending, a 13% reduction in government-spending volatility relative to countries without such funds, reduced fiscal procyclicality among 42 oil-exporting countries (especially where institutions are strong), and decreased volatility in exchange rates, inflation, and money circulation. But this pattern is not universal: one study conditioning on institutional quality finds no statistically significant effect on fiscal policy outcomes, another finds that funds exacerbate spending volatility specifically where governments are otherwise fiscally imprudent, and a third finds the direction and size of the effect depends on changes to the fund's own management rules. A related, narrower literature on fiscal-performance measures (resource balance, debt, non-resource exports) splits the same way: one study conditioning on institutional quality finds a positive effect, another conditioning on the same variable finds none.

Where funds actually pay out money, the evidence is descriptive rather than causal, but the magnitudes are concrete. Alaska's Permanent Fund dividend has accounted for up to 5% of per capita annual income for Alaskans generally, and up to 10% for the state's Indigenous populations. New Mexico's permanent-fund transfers are estimated to lighten the average household's tax burden by around $1,000 annually. These funds also operate under fixed, publicly stated payout formulas: Alaska reinvests 50% of fund earnings and distributes the other 50% as dividends; New Mexico withdraws 5% of its Land Grant and Severance Tax Permanent Funds annually, earmarked for public education; and Norway lowered its own budgetary spending rule from 4% to 3% of fund value in 2017.

The strongest causal evidence on payout-rule behavior comes from university endowments, which follow a structurally similar model (a payout rate applied to a multi-year moving average of fund value) and show a clear asymmetry: endowments follow their own stated payout policies after positive portfolio shocks, but actively cut payouts below the level their policies prescribe after negative shocks. A negative shock equal to 10% of a university's budget reduces the payout rate by 17–19 basis points; a 10% negative portfolio return reduces the payout rate by 10.8%, or about 0.56 percentage points off an average payout rate of 5.2%. Within a subsample with more detailed policy data, a 10% negative shock produces an even larger deviation, 47–57 basis points below what the stated policy would prescribe. Endowments do not show this same responsiveness to shocks in other revenue sources such as government appropriations or research grants — that relationship is not statistically distinguishable from zero.

Separately, endowments revise their permanent spending rules often — a weighted mean of almost 25% of endowments changed policy in a given year between 2003 and 2011, and half of a large sample changed policy between one and eight times over that period. The likelihood of revision is associated with larger endowment size, lower historical portfolio returns, and lower historical payout levels; it is not associated with how large a share of the institution's operating budget the endowment funds. Across the same sample, Moving Average payout formulas are the dominant design (used by 65–76% of endowments each year) and are retained longer than simpler percentage-of-income rules. Spending-rule changes tend to precede asset-allocation changes, but endowments that revise their spending rules show no measured difference in benchmark-adjusted investment performance from those that do not. After the 2008 financial crisis, endowments became more likely to preserve their existing permanent policies despite funding pressure, while about one-third adopted temporary supplemental payouts on top of those permanent rules; one further source reports that universities increased their allocations to illiquid, nontransparent assets after 2008 relative to before it, though without a stated magnitude.

A separate strand covers investment-policy design rather than payout behavior. Simulation work finds that a 1-percentage-point increase in investment returns, from 3% to 4%, could compound into reserves roughly 11 times larger by 2100. Cross-pension governance research finds that adding one retired member to a fund's governing board decreases funding by 1.7 percentage points, and adding one active-worker member reduces stock funding by 0.7 percentage points, while enhanced financial reporting to stakeholders is associated with a 2.1-percentage-point increase in annual returns; governance variables more broadly account for over 20% of cross-plan variation in investment strategy, which is itself associated with subsequent return performance. A separate optimization exercise finds that, given a 20%-of-salary contribution rate in one modeled scenario, the return-maximizing allocation is 30% equities and 70% bonds — a result specific to that scenario's assumptions. On real-world fund design, comparative work on Alaska and Norway finds that dedicating revenue to a fund under constitutional protection restricts government manipulation of contribution rules and is associated with consistent fund growth, that pegging payouts to the fund's expected real rate of return protects principal at the cost of short-term fiscal flexibility, and that tight withdrawal-rule controls are associated with protecting fund principal from being redirected when political or economic conditions change; conversely, integrating a savings-focused fund into the normal government budget process is associated with increased short-term spending at the expense of long-term fund objectives. Norway's Government Pension Fund–Global also operates under explicit statutory limits: equities held at 30–50% of the portfolio and bonds at 70–50%, no more than 3% of any single company's voting shares or share capital, and diversification enforced across seven distinct levels (asset class, region, currency, country, industry, company, and issuer type); Russia's National Wealth Fund is authorized to hold a defined set of foreign-currency and ruble-denominated eligible assets. A theoretical optimum derived from simulation, distinct from any real fund's actual rule, finds that optimal spending should be procyclical and lower than the expected financial return, that the optimal equity share should itself be procyclical and rise from about 25% of the portfolio at low wealth to about 75% at high wealth, that spending cuts following losses should be larger than proportional to the wealth loss, that spending should be kept particularly low at low wealth levels to rebuild reserves, and that simple rules with a constant equity share and payout ratio can reach nearly the same expected utility as the fully optimal nonlinear policy.

No claim in this evidence set measures the effect of permanent-fund payout rules on broader outcomes such as regional economic growth, poverty, or labor supply — the coverage here is limited to fiscal volatility, portfolio/payout mechanics, and fund governance.

Where the evidence splits

Resource-revenue funds' effect on government spending volatility and procyclicality is a three-way split. Several studies find funds reduce volatility or procyclicality — a 13% reduction in spending volatility, a smoothing effect on spending, reduced procyclicality among 42 oil-exporting countries, and reduced volatility in exchange rates, inflation, and money circulation. One study, conditioning its estimates on institutional quality specifically to address endogeneity, finds no statistically significant effect at all. Another, using a pooled cross-sectional approach, finds the opposite sign: funds exacerbate spending volatility, but specifically among governments assessed as otherwise fiscally imprudent — funds free up more resources for governments already prone to volatile spending. A fourth study attributes any remaining heterogeneity to changes in the funds' own management rules rather than to funds as such. The sources state the moderating condition for the reversal (government prudence) and for the null result (conditioning on institutional quality), but they do not reconcile why conditioning on institutional quality produces a null result in one study and a positive effect in a related study on fiscal-performance measures below.

Resource funds' effect on broader fiscal-performance measures (nonrenewable-resource balance, debt, non-resource exports) also splits. One study, conditioning on institutional quality, finds a positive effect on these measures. Another, also conditioning on institutional quality, finds no statistically significant effect on fiscal policy outcomes. Both use a similar identification strategy (controlling for institutional quality to address endogeneity) and reach different conclusions; the claims here do not explain the discrepancy beyond noting it exists, so it is unexplained by the sources on this page.

Thresholds and dose-response

Endowment payout behavior changes sign at a zero-return threshold: following positive portfolio shocks, endowments follow their own stated payout policy with no measured deviation; following negative shocks, they actively cut payouts below the policy-implied level. The size of that cut scales with the shock: a negative shock equal to 10% of a university's budget reduces the payout rate by 17–19 basis points; a 10% negative portfolio return reduces the payout rate by 10.8%, or about 0.56 percentage points relative to an average payout rate of 5.2%; within a subsample with more detailed policy data, a 10% negative shock produces a 47–57 basis point reduction relative to what the stated policy would prescribe.

A 1-percentage-point increase in investment returns, from 3% to 4%, is associated with reserves approximately 11 times larger by the year 2100 in one simulation. Each additional retired member added to a pension governing board is associated with a 1.7-percentage-point decrease in funding; each additional active-worker board member is associated with a 0.7-percentage-point decrease in stock funding. In a separate optimization simulation, the optimal equity share rises with portfolio wealth, from about 25% of the portfolio at low wealth levels to about 75% at high wealth levels. Norway's Government Pension Fund–Global operates under a statutory band holding equities between 30% and 50% of the portfolio (bonds the complementary 70–50%) and caps investment in any single company at 3% of that company's voting shares or share capital. Norway also lowered its own budgetary spending rule from 4% to 3% of fund value in 2017, a one-percentage-point reduction in the statutory withdrawal rate.

Open empirical questions

• Do the resource-fund fiscal-volatility findings and the endowment payout-asymmetry findings hold for state-level permanent trust funds specifically, which have different governance and political-accountability structures than either sovereign wealth funds or university endowments?
• What explains the disagreement between studies that condition on institutional quality and reach opposite conclusions on resource funds' effect on fiscal-performance measures (positive versus null)?
• What is the causal effect of a specific statutory payout rate (for example, 3% versus 5% versus 7% of fund value) on long-run fund depletion risk, holding return assumptions constant — the sources describe several funds' chosen rates but do not measure the consequence of choosing one rate over another?
• Do dividend-based permanent funds (Alaska-style universal payments) show different political durability or "raid" resistance than earmarked-transfer funds (New Mexico-style) in a direct, controlled comparison?