The Investigative Journal’s weekly survey of policy research tracks what Washington’s think tanks published over the past week, summarizes their findings, and notes each institution’s political orientation and major sources of funding so readers can weigh the analysis accordingly. This week’s reports span the U.S.-China military balance, the regulation of open-weight artificial intelligence, the politics of balanced budgets, the running cost of the 2026 tariffs, the economic value of trustworthy government data, wage policy, and the fiscal effect of immigration.
RAND maps how the U.S.-China military gap shifted from 2017 to 2024
The RAND Corporation on May 12 published “Key Changes in U.S. and Chinese Military Capabilities, 2017–2024,” a reference paper by Jacob L. Heim and Cristina L. Garafola that compiles open-source data on how the forces of the United States and the People’s Republic of China have evolved over seven years. The authors note that U.S. defense policymakers identify China as the most formidable military competitor the United States faces, and the paper is intended as a non-technical resource for officials and the public following that competition.
Rather than offering a single verdict, the paper assembles comparative data across categories of military hardware and capacity, extending the kind of analysis RAND has produced in earlier “scorecard” work. The study was commissioned by the Secretary of the Air Force and conducted within RAND Project AIR FORCE.
RAND is a nonprofit, nonpartisan research organization founded in 1948 with roots in U.S. Air Force–sponsored research. It operates several federally funded research and development centers, and a substantial share of its budget comes from U.S. government contracts; it also draws on philanthropic gifts and grants. Readers should note that this particular paper was government-commissioned, which RAND discloses on the publication page.
RAND argues open-weight AI models need their own evaluation playbook
In a separate May 4 release, “Open-Weight AI Models Require Proportional Evaluation Approaches,” RAND researchers Patricia Paskov, Christopher Rodriguez, Sunishchal Dev, and Stephen Casper contend that the safety-evaluation practices built for closed, vendor-controlled AI systems do not map cleanly onto open-weight models, whose parameters can be freely downloaded, modified, and redistributed.
The authors propose a set of “proportional evaluation” criteria and then review the practices applied to 37 families of open-weight models released between early 2025 and April 2026. Their finding is pointed: only one of the 37 families met all of the proposed criteria, and most met none. The paper frames this as a gap between how quickly open-weight models are proliferating and how thoroughly they are being assessed for risk.
This report was produced by RAND’s Center on AI, Security, and Technology within its Global and Emerging Risks division, and RAND states it was independently initiated and funded through income from operations and philanthropic gifts and grants rather than a government client.
Brookings: the late-1990s balanced budgets were mostly an “accident”
Writing for the Urban-Brookings Tax Policy Center on May 11, fellow Jessica Riedl published “How did the budget get balanced in the late 1990s?” as part of a series drawn from her 2026 federal budget chart book. Riedl’s argument is that the budget surpluses of 1998 through 2001 were largely a “temporary historical accident” rather than a product of the budget deals associated with President Bill Clinton and House Speaker Newt Gingrich.
Using data from the Congressional Budget Office, the Office of Management and Budget, and the Treasury Department, the analysis attributes roughly a quarter of the deficit reduction between 1992 and 2000 to post–Cold War defense cuts — defense spending fell to 2.9 percent of GDP, its lowest share since the 1930s — and roughly 60 percent to a temporary stock-market and capital-gains tax revenue boom. Clinton’s 1993 tax increases, the piece estimates, accounted for about one-tenth of the improvement. When the dot-com bubble burst and the September 11 attacks reversed defense reductions, the surpluses quickly disappeared.
The Brookings Institution is one of Washington’s oldest think tanks and is generally described as centrist to center-left; it states it is supported by a diverse array of foundations, corporations, and governments and that each publication reflects only its authors’ views. Riedl is notable as a fiscal-conservative budget analyst, an illustration that institutional labels do not always predict an individual author’s leanings.
Tax Foundation keeps a running tally on the 2026 tariffs
The Tax Foundation’s continuously updated “Tariff Tracker: 2026 Trump Tariffs & Trade War by the Numbers,” maintained by Erica York and Alex Durante, was refreshed again on May 5. The tracker estimates that the tariffs imposed during the trade war amount to an average tax increase of roughly $700 per U.S. household in 2026 and concludes that, to date, they have not meaningfully altered the U.S. trade balance.
The tracker’s timeline also records a significant legal development: a Supreme Court ruling that the president cannot impose tariffs under the International Emergency Economic Powers Act, which the Tax Foundation has folded into its revenue and economic estimates. Because the tracker is revised as policy changes, its headline figures shift over time; readers should treat the numbers as a snapshot tied to the May 5 update.
The Tax Foundation is a nonprofit founded in 1937 that describes its mission in terms of pro-growth tax policy; it is generally regarded as free-market or center-right in orientation and is funded by foundations and individual donors.
AEI scholars put a dollar value on trustworthy government statistics
A working paper released May 4 through the National Bureau of Economic Research, “The Value of Reliable Statistics,” co-authored by Stanford economist Nicholas Bloom, former Bureau of Labor Statistics commissioner Erica Groshen, and American Enterprise Institute economists Duncan Hobbs and Michael R. Strain, attempts to quantify what is lost when confidence in official data erodes.
The authors use an event-study design centered on the August 1, 2025 firing of the head of the Bureau of Labor Statistics to estimate the resulting increase in policy uncertainty, then translate that into macroeconomic terms. Their baseline estimate is that preserving trust in the integrity of official statistics generates roughly $25 in economic benefits for every $1 spent on a statistical agency’s budget — a finding the authors argue should inform debates over the funding and independence of federal data agencies.
AEI is a center-right think tank founded in 1938; it is funded by foundations, corporations, and individual donors. The paper itself was distributed by NBER, an independent nonpartisan research organization.
Center for American Progress pitches “sectoral pay standards”
On May 5, the Center for American Progress published “How Market-Based Sectoral Pay Standards Raise Wages and Improve Affordability,” by senior fellow Karla Walter. The report argues that policymakers can raise wages and ease affordability pressures by extending prevailing-wage rules and similar pay standards — long used in public contracting — into parts of the private sector.
Walter’s framing is that such standards can be designed to track local market rates rather than impose flat mandates, which the report contends would lift pay in lower-wage industries while limiting disruption to employers. The piece builds on a body of CAP work on prevailing wages and industry standards boards.
The Center for American Progress is a progressive, center-left think tank founded in 2003; it is funded by foundations, corporations, and individual donors, and has historically received support from labor organizations. Its policy recommendations should be read in that context.
Cato Institute tallies the fiscal effect of immigration
Rounding out the week is the Cato Institute’s white paper “Immigrants’ Recent Effects on Government Budgets: 1994–2023,” which examines three decades of tax and spending data. The study concludes that immigrants — both legal and unauthorized — generated substantially more in federal, state, and local taxes than they induced in government spending, and that immigration reduced cumulative budget deficits by about a third over the period, with the annual fiscal benefit reaching an estimated $878 billion in 2023.
The paper reports that immigrants accounted for about 14 percent of tax revenue but only 7 percent of government spending across the period, and estimates that without immigration, U.S. public debt would stand far higher as a share of GDP than its actual 2023 level. As with any single study, the findings rest on the authors’ modeling assumptions, and other analysts using different methods have reached more mixed conclusions on immigration’s fiscal effects.
The Cato Institute is a libertarian think tank founded in 1977 with early support from Charles Koch; it is funded today by foundations and individual donors and consistently favors looser immigration and freer markets.
Why this matters for TIJ’s beats
Several of this week’s reports speak directly to the accountability questions The Investigative Journal follows. The AEI-affiliated paper on data integrity bears on the reliability of the official statistics that underpin oversight reporting. The Brookings, Tax Foundation, and Cato analyses each touch federal fiscal policy — how deficits are built, what tariffs cost households, and how demographic change shapes the budget. And RAND’s defense work informs scrutiny of national-security spending and strategy. Read together, and read against one another, they offer a cross-spectrum picture of the evidence policymakers are working from. As always, readers are encouraged to consult the original reports, linked above, and to weigh each institution’s orientation and funding when assessing its conclusions.