Measuring the law
How much law is on the books, and is there more of it than there used to be? Counting laws passed each year cannot answer that, because most laws amend laws that already exist. You have to measure the code itself. Do that and federal law turns out to have grown by nearly half since 2000 — over the same quarter century in which Congress passed steadily fewer laws.
Counting laws passed each year cannot tell you how much law exists, because most laws amend laws that are already there. The way around that is to stop counting enactments and measure the code itself — the consolidated text, with every amendment and repeal already folded in. Do that at two dates and the amendments net out on their own. For federal law the record goes back a century.
The United States Code, 1926 to 2023
Words of federal statute in force in each published edition. Editions appear every six years through 1988 and annually from 1994; 1989 to 1993 were never published in this dataset and are left out rather than drawn through.
This is the measure you want if the question is "how much law is there." It is a count of words in the codified text, so an amendment that rewrites a section changes the number by exactly the amount of text it adds or removes, and a repeal reduces it. Because the counted text includes section headers and historical notes, the absolute level runs roughly 54 percent above counts that cover operative statutory language only, so the trajectory is more reliable than the raw total. Editions through 1988 were scanned from print and are lower confidence than the 1994-onward figures, which came from official electronic text.
This chart sets three measures of federal law in force and one measure of federal law enacted per year side by side, all indexed to 2000. The stock lines rise by 41 to 46 percent. The flow line falls by two thirds. Congress has been passing fewer laws for a quarter century and the body of federal law has grown by nearly half regardless.
Federal law in force against federal laws enacted, indexed to 2000 = 100
Solid lines are stock, the size of the law on the books. The dashed line is flow, public laws enacted that year.
The two US Code lines are independent measurements of the same thing by different people using different units — words counted from a research dataset, pages counted from government publication metadata. They land on 146.1 and 146.3. Nothing about the two methods was coordinated, so that agreement is evidence that the stock measure is sound.
New York's Consolidated Laws can be retrieved as they stood in any week since October 2014, which makes it the one government besides Washington whose code can be measured directly. Both series below count the same object in the same unit: words of statute in force. Over the eight years where both exist, federal statutory law grew roughly twice as fast as New York state's.
Law in force, indexed to 2015 = 100
2015 is the earliest year all three can be measured, because New York's archive begins in October 2014. Each series ends where its source ends.
Over 2015 to 2023: federal statute in force rose 14.2 percent, New York state statute 7.4 percent, the federal regulatory code 6.9 percent. New York's full run to 2026 is 11.0 percent, about 0.95 percent a year. The absolute levels are not comparable across the two statutory series — the federal word count includes headers and historical notes and the New York count does not — but the growth rates are, because each is measured against its own starting point.
This series passes the checks the New York City attempt failed, which is the reason it is plotted and the city is not. Growth computed across all laws and across a fixed cohort of the 133 laws present in every snapshot agree to within 0.005. The number of laws in the corpus barely moves, 133 to 135 in eleven years. Year-over-year change is smooth, between +0.21 and +2.12 percent, with no step between adjacent years. And the API correctly returns "not found" for the Cannabis Law before its enactment in 2021, which is independent evidence that it is serving real historical versions rather than today's text with an old date attached.
Each body passes a different kind of law in different quantities, so the levels cannot be compared directly. Setting each to 100 in 2000 shows the shape of the change instead. The City Council pulls away; Albany oscillates around its starting point; Congress drifts down.
Laws enacted per year, indexed to 2000 = 100
Higher means more laws passed that year than in 2000. New York City ends at 2024, the last complete year in its source.
Indexing to a single year makes the comparison sensitive to that year. 2000 was an ordinary year for all three bodies, but New York City's 2000 figure of 74 local laws is a small base, so its index moves further on the same absolute change. The raw counts are directly below, and the table above gives every underlying number.
New York state passes several hundred chapter laws a year, most of them narrow. New York City passes dozens to a couple hundred local laws. Congress passes one to three hundred public laws. Note the vertical scales differ; each panel is drawn to its own range.
New York City's single-year spikes line up with the final year of each Council term, when bills pending since the start of the session are cleared: 2005, 2009, 2013, 2017 and 2021 are all term-end years, and each sits above both of the years on either side of it. The level of the whole series also shifts upward after 2015, so recent years outside the term-end cycle are well above the early 2000s too. New York state's 2020 trough of 389 chapter laws is its lowest figure since at least 1995.
The counts above are flow — what gets enacted in a given year. A legal code is stock — everything enacted and not yet repealed. Flow can fall for years while stock keeps climbing. The Code of Federal Regulations shows how: Congress passed 37 percent fewer laws per year, and the regulatory code still grew by 56,346 pages. Stock can shrink, though. Measured back to 1950, the regulatory code has shrunk in 15 of 74 years, most sharply by 6,074 pages in 1996 and 5,895 in 1985. It falls less often than it rises, and by less when it does.
Stock: pages in the Code of Federal Regulations
The accumulated size of the federal regulatory code, measured each year. Ends at 2024.
Flow: pages published in the Federal Register
What was added to the record in each year, including proposals and notices that never become code.
The two panels are separate charts because they are different measures on different scales. Overlaid on twin axes, they would imply a relationship produced by the choice of axes. The Code of Federal Regulations series ends at 2024; no 2025 edition has been published as of retrieval.
Federal Register pages fell 43 percent between 2024 and 2025, from 106,109 to 60,917. Breaking the total into its four document categories shows the fall was not spread evenly. Pages of final rules dropped 63 percent. Pages of presidential documents went the other way, to 2,072 — the highest in any year since this category table begins in 1976, and well clear of the previous high of 1,301 in 1995.
Federal Register pages by document category
Stacked; the four categories sum to the annual total.
Presidential documents, on their own scale
The green band above is too thin to read in the stacked chart. Drawn alone, at roughly one fiftieth of the vertical scale, 2025 separates from every other year in the series.
Presidential documents are a small share of the total in every year — 2,072 pages against 60,917 in 2025 — so the record is a record within a narrow category, not a claim that presidential action replaced rulemaking in volume. A page count also says nothing about the legal reach of what is on the page. 2026 is excluded throughout as an incomplete year.
Counting laws is not the same as measuring how much law there is. These are the limits of what this page shows.
A measurement that was made, checked and thrown away
The New York City Administrative Code is distributed as bulk XML, and the Internet Archive holds copies of that file from 2022, 2023, 2025 and 2026. Parsing all of them produced an apparent increase from 9,656 to 12,206 sections in four years — growth of 26 percent.
The growth is an artifact. Sections, words, chapters and the publisher's own file count all grew by almost exactly the same factor, which is the signature of a distribution getting bigger rather than a body of law. In the appendix of unconsolidated local laws, the earliest law present is 1985 in every snapshot from December 2023 onward, but 2015 in the June 2022 snapshot. The publisher backfilled three decades of local laws in a single step. Those 1,332 laws were on the books in 2022 and appear as new sections only because the earlier file did not list them.
Of the apparent growth, 52 percent is provable publisher backfill, 17 percent is unnumbered technical material that moves in the same step, and at most 30 percent reflects law actually enacted between 2022 and 2026. Restricting to numbered code sections gives a defensible but short and still-imperfect result — 7,965 to 8,590 sections, up 7.8 percent over four years, with an estimated 11 percent residual contamination. That is roughly 2 percent a year, which is what you would expect of a city passing about 150 local laws a year that mostly amend what is already there. It is not plotted here, because four years of a partly contaminated measure does not belong on the same page as a century of federal data.
Every one of these limits, the sources, the queries and the checks that failed are documented in full in the methodology. The underlying data is in the repository as JSON, one file per source, with provenance on each.