Written and maintained by CASRAI Editorial Board
Published
On this page
A Statement of Administration Policy (SAP) is a written statement of the executive branch’s position on a specific bill, issued by the Office of Management and Budget (OMB) for the Executive Office of the President. SAPs run from strong support to strong opposition and include veto threats, which makes them a ready-made, dated record that political scientists code and count when they study how presidents bargain with Congress. This guide explains what a SAP is, how to read one, how researchers have counted them, and a worked example using the published Bush and Obama counts.
What a Statement of Administration Policy is
The Congressional Research Service (CRS) describes SAPs as the administration’s formal communication to Congress about pending legislation. The White House decides whether to issue one for a given bill; career civil servants draft the text together with political appointees, and OMB releases it. SAPs grew out of OMB’s tracking of appropriations bills and expanded in use under President Reagan. Modern SAPs are published online and sent to all Members of Congress, and since 2012 the White House has also announced them on the @OMBPress account.
A SAP is not law, not an executive order and not a signing statement. A signing statement is issued after a bill passes; a SAP is issued while the bill is still moving, so it is a signal to legislators rather than a legal act. See also the CASRAI entry on congressionally directed funding for one place where the appropriations process meets executive positions.
Anatomy of a SAP
According to CRS, a SAP typically gives the bill title, its sponsor and the number of cosponsors, then the administration’s position and an operative clause of support or opposition. It may go on to name specific provisions and suggest revisions. Some drafts carry internal text, set off below a row of stars, that is not part of the public version. Placement of the position varies: SAPs under Presidents Bush and Clinton stated it in the first paragraph, while those under President Obama often put it in the middle or at the end. A coder therefore cannot assume that the first paragraph holds the position.
When SAPs are issued
SAPs are usually released when a bill reaches the House Rules Committee or the floor. Most bills receive one SAP, but some administrations issued several as the text changed. This matters for measurement: the number of SAPs is not the number of bills.
Veto-threat language in SAPs
CRS identifies two standard formulations. In the first, the SAP says that if the President were presented with the bill, he would veto it. In the second, the President’s senior advisors would recommend that he veto the bill. CRS notes that the first formulation may leave the administration less room to negotiate. Since the 108th Congress the veto-threat sentence has been underlined in the published SAP, which helps when scanning documents by eye or by script.
How researchers count and code SAPs
CRS report R46338 on veto threats in the George W. Bush and Obama administrations is a useful template for transparent counting rules. Its choices were:
- Population: only SAPs on non-appropriations bills were analysed.
- Unit: each SAP is treated as one veto threat. If one bill drew several SAPs with threats, each was counted separately, so a bill can appear more than once.
- Coding: a threat is attributed either to the President or to senior advisors, according to the formulation above.
- Outcomes: bill status was taken from Congress.gov. Bills later folded into other legislative vehicles were not tracked.
- Vetoes: regular and pocket vetoes are not distinguished.
Report the rule you adopt for each of these. A different unit (per bill rather than per SAP) or a different population (including appropriations) will change every rate below. The CASRAI entry on operationalizing variables covers the general logic of turning a document feature into a measured variable.
Worked example: veto-threat share and outcomes
R46338 reports 491 non-appropriations SAPs for the George W. Bush administration and 472 for the Obama administration. Veto threats appeared in 24 presidential and 94 senior-advisor SAPs under Bush, and 43 presidential and 186 senior-advisor SAPs under Obama.
Share of SAPs carrying a veto threat.
- Bush: (24 + 94) / 491 = 118 / 491 = 0.2403, or 24.0%.
- Obama: (43 + 186) / 472 = 229 / 472 = 0.4852, or 48.5%.
- Both combined: (118 + 229) / (491 + 472) = 347 / 963 = 0.3603, or 36.0%.
Veto rate among bills that passed both chambers. Pooling both administrations, 17 bills that passed after a presidential threat were either signed (1 + 4 = 5) or vetoed (7 + 5 = 12), so 12 / 17 = 70.6% were vetoed. After a senior-advisor threat, 36 passed bills were signed (16 + 17 = 33) or vetoed (1 + 2 = 3), so 3 / 36 = 8.3% were vetoed. These two percentages match the ones CRS reports.
The data table below sets out the four threat groups, with a computed veto share among threatened bills that reached the President’s desk.
Interpreting the numbers
| Observation | What it supports | What it does not support |
|---|---|---|
| Share of SAPs with a threat rose from 24.0% to 48.5% | Threats appeared in a larger fraction of published SAPs in the Obama sample | A claim that Obama threatened more bills in total without checking how many bills each SAP count represents |
| Most threatened bills never reached the desk | Many threatened bills stalled in Congress (for example 77 of 94 advisor-threat bills under Bush) | That the threat caused the stall; CRS says the method makes direct effects difficult to determine |
| Veto share is far higher after a presidential threat | The two formulations are associated with very different outcomes among passed bills | A forecast for any single bill; the rates rest on 17 and 36 bills |
CRS cautions that general trends may be observed in aggregate, but that the sample is small, with 12 vetoes in each administration, and that the method cannot isolate the direct effect of a threat on a bill’s outcome.
Assumptions and common mistakes
- Treating SAPs as bills. Multiple SAPs on one bill inflate counts unless you deduplicate or state the per-SAP rule.
- Mixing populations. R46338 excludes appropriations SAPs. Comparing its figures with a count that includes them gives a mismatch.
- Reading a threat as a cause. Bills that never passed both chambers may have died for reasons unrelated to the SAP.
- Losing bills to other vehicles. A provision moved into a different bill disappears from a status-based outcome count.
- Using percentages without denominators. Always report the base, for example 12 of 17, not only 70.6%.
- Assuming a fixed layout. The position can sit in the first paragraph or near the end, so coding rules should read the full text.
How to report and cite SAPs in research
Record, for every SAP in your dataset, the bill number and title, the SAP’s date, the issuing office (OMB, Executive Office of the President), the coded position, the veto-threat type if any, the URL of the copy you coded and the archive it came from. Name your counting rules in the methods section and give the denominators for each rate. When you reuse the CRS figures, cite the report by number and year rather than recomputing from memory; when you build your own counts, deposit the coding sheet and a codebook such as one structured along the lines of the DDI Codebook so others can reproduce the coding.
Where to find SAPs
CRS lists these sources: the OMB page on the White House website (current administration only), archived White House websites for earlier administrations, the American Presidency Project at UC Santa Barbara, and the @OMBPress account. R46338 also names the Internet Archive and Congress.gov as sources for the data behind its analysis.
Calculating the share in Excel, R or Python
Excel, with presidential threats in B2, advisor threats in C2 and total SAPs in D2:
=(B2+C2)/D2
R:
threats = c(bush = 24 + 94, obama = 43 + 186)
saps = c(bush = 491, obama = 472)
round(100 * threats / saps, 1) # 24.0 48.5
Python:
threats = {"bush": 24 + 94, "obama": 43 + 186}
saps = {"bush": 491, "obama": 472}
print({k: round(100 * threats[k] / saps[k], 1) for k in saps}) # {'bush': 24.0, 'obama': 48.5}
For basic summaries of this kind see the CASRAI guide to descriptive statistics in Excel.
Sources
- Congressional Research Service, R44539, Statements of Administration Policy (2016)
- Congressional Research Service, R46338, Veto Threats and Vetoes in the George W. Bush and Obama Administrations (2020)
- White House, Office of Management and Budget: Statements of Administration Policy
- The American Presidency Project, University of California, Santa Barbara: Statements of Administration Policy
Data: Veto-threat SAPs by threat type and bill outcome, Bush and Obama administrations (non-appropriations)
Counts of non-appropriations Statements of Administration Policy carrying a veto threat, split by presidential versus senior-advisor threat, with the bill outcome for each group as reported by CRS R46338; the last column is computed as vetoed divided by signed plus vetoed.
Period covered: 2001/2017. Coverage: United States.
| Administration and threat type | Threat SAPs (count) | Signed (bills) | Vetoed (bills) | Did not reach President (bills) | Veto share of bills reaching desk (percent) |
|---|---|---|---|---|---|
| Bush, presidential threat | 24 | 1 | 7 | 16 | 87.5 |
| Bush, senior-advisor threat | 94 | 16 | 1 | 77 | 5.9 |
| Obama, presidential threat | 43 | 4 | 5 | 34 | 55.6 |
| Obama, senior-advisor threat | 186 | 17 | 2 | 167 | 10.5 |
Sources
- CRS R46338, Veto Threats and Vetoes in the George W. Bush and Obama Administrations (accessed 2026-11-18)
Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.







