IMLS Library Visits vs E-Circulation: Visits Index 67, E-Circ 153 — and 7 States Where Both Fell

Public libraries are often narrated as either empty buildings or digital triumphs. The Institute of Museum and Library Services (IMLS) Public Libraries Survey (PLS) lets us test a sharper claim: did physical visits fall while electronic circulation rose, and where did both channels drop together?

Indexed to FY 2019 (=100), the national desk reading for FY 2023 is stark. Visits sit near index 67 — a recovery from pandemic troughs, but still far below the pre-COVID baseline of about 1.2 billion annual visits. Electronic circulation sits near index 153, with benchmarking means around 1.68 items per person versus 1.1 in FY 2019. Median e-circ per person in the IMLS FY 2023 research brief rose 84% from FY 2018 (0.49 → 0.90). The dual-track chart in the dashboard above is the national story in one frame: a deep visit collapse and partial rebound on one axis; a steady digital climb on the other.

A 51-jurisdiction desk panel (50 states + D.C.) classifies each place by those two indexes. Thirty-eight land on the substitution path — visits still below FY 2019, e-circ above it. Seven are dual-decline: both indexes below 100. Only a handful have restored visits at or above the 2019 baseline. That geography is the policy-relevant half of the question. National averages hide places where digital growth never offset the walk-in loss — and places where neither channel has come back.

What the PLS actually measures

The PLS is a near-census of U.S. public library systems, collected with state library agencies. Core service fields include library visits, physical item circulation, and electronic materials circulation (e-books, downloadable audio, and related e-video), plus hours, programs, and staffing. Per-person rates use the population of the legal service area so large and small systems can be compared.

Two measurement notes matter for this post. First, medians and means diverge. The FY 2023 research brief emphasizes medians (overall median visits per person 2.8 in FY 2023, up from 0.4 in FY 2021; median e-circ 0.90). Benchmarking tables and older findings reports often quote national means and totals — e.g., FY 2019’s 1.2 billion visits and 1.1 electronic items circulated per person. Our indexes blend those disclosed anchors into a FY 2019 = 100 frame; they are desk composites, not a single IMLS published index series. Second, fiscal years are not calendar years. Libraries report on state-defined fiscal calendars; pandemic effects land in different PLS years depending on when a system’s year closed. FY 2020–FY 2021 are the trough for visits; FY 2022–FY 2023 are the rebound years.

Caveat: visit counts can include gate and door counter methods that differ across systems; electronic circulation depends on vendor reporting and how consortia attribute downloads. Cross-year comparisons are best read as directional, especially for small states.

The national dual track: collapse, rebound, and a permanent gap

From FY 2018 through FY 2019, visits were high and e-circ was already rising from a low base. Then buildings closed. IMLS reports that 90% of libraries temporarily closed buildings due to COVID-19 in FY 2020; that share fell to 15% by FY 2022. Median visits per person cratered to 0.4 in FY 2021 before climbing to 2.5 (FY 2022) and 2.8 (FY 2023). Hours open largely returned toward pre-pandemic levels by FY 2023, especially in cities.

Physical circulation followed a related path: down in the trough, then back toward roughly 3.91 median items per person in FY 2023. Electronic circulation did not wait for reopenings. It rose every year in the brief’s window — suburban medians from 0.63 to 1.26, city medians from 0.69 to 1.31, with town and rural systems growing more slowly but still up.

The economic reading is not “libraries are dying” or “libraries went fully digital.” It is path dependence with a gap. Door traffic remains structurally below 2019 even after a strong two-year recovery. Digital borrowing is structurally above 2019 and still climbing. The 86-point gap between visits index (~67) and e-circ index (~153) is the headline tension boards and state librarians are managing: buildings, staffing, and programs sized for in-person demand, plus license and platform costs sized for digital demand that did not reverse when doors reopened.

State quadrants: substitution, dual decline, and rare full recovery

Classify each state by whether its FY 2023 visits index and e-circ index clear 100. Four cells appear:

The substitution path dominates — as the national indexes predict. Large coastal and Sun Belt systems (California, New York, Florida, Texas, Massachusetts, Washington, Oregon) combine soft visit indexes with strong e-circ indexes. That is consistent with dense broadband markets, OverDrive/Libby-style platforms, and patrons who learned remote borrowing in 2020 and kept the habit.

The dual-decline cell is smaller but more alarming. In the desk panel it includes Mississippi, West Virginia, Louisiana, New Mexico, Alabama, Arkansas, and Kentucky. These are not identical library systems — some are rural-heavy, some face thin operating revenue per capita historically — but they share a pattern: visits remain deep below 2019 and electronic circulation never cleared the pre-pandemic bar on the index. Digital did not automatically substitute for closed buildings or reduced hours. Where connectivity, device access, or e-collection budgets are weak, the pandemic shock can leave both channels damaged.

Full recovery is rare and concentrated in smaller Western and Plains states in this panel (Utah, Wyoming, North Dakota, South Dakota) — places where visit culture was already intense per capita and where the absolute digital market is smaller. In-person-only outliers (Alaska, Montana in the desk cut) restored visits while e-circ indexes still lag — a reminder that broadband and platform uptake are uneven even when buildings fill again.

Locale: cities and suburbs lead the digital climb

IMLS locale codes (city, suburb, town, rural) show the digital gradient clearly. City and suburban systems posted the largest relative e-circ gains from FY 2018 to FY 2023. Rural e-circ in FY 2023 remained roughly half the city median rate in the brief’s framing. Visits recovered across all locales from the FY 2021 floor, but rural and town medians remain lower than city/suburb medians on both visits and e-materials.

For public finance, that matters. State aid formulas and local millages often treat “library service” as a single good. The PLS says it is at least two goods with different geographies: in-building civic space and licensed digital content. Suburban systems can look “healthy” on e-circ while still fighting for visit-driven program attendance; rural systems can look “quiet” on apps while remaining essential physical infrastructure for homework, printers, and cooling centers.

Dual decline is not the same as low use

A dual-decline label is about change versus FY 2019, not about whether a state has “good” libraries. Some dual-decline states still circulate materials and welcome visitors at meaningful absolute rates; they simply have not regained the 2019 combination of door counts and downloads. Conversely, a substitution-path state with visits index 60 and e-circ index 170 may be pouring money into e-content licenses while branch hours stay constrained.

The dashboard’s dual-decline bars and regional stacks make the fiscal geography visible: the South carries most dual-decline counts in this panel; the Northeast and West are heavily substitution-path; the Midwest mixes strong Plains recoveries with industrial-state substitution. Policy that only celebrates national e-circ growth will miss the dual-decline cluster. Policy that only mourns visit losses will miss how thoroughly digital demand has rebased upward in high-connectivity markets.

Caveats, confidence, and what this desk does not claim

State indexes are desk composites from PLS state ranking and services tables across overlapping years, not an official IMLS “visits vs e-circ index” product. Small states are noisier; D.C. is an outlier jurisdiction; imputation and nonresponse in the PLS, while historically low, still exist. Electronic circulation definitions have matured since IMLS began reporting the series more systematically around FY 2018, so early-year levels are less comparable than 2019→2023 changes.

This post does not estimate welfare — whether a download substitutes for a visit one-for-one, or whether program attendance and Wi-Fi sessions should count as “use” alongside checkouts. It does not price OverDrive contracts or staff FTE against circulation mixes. It asks a narrower empirical question and answers it with disclosed IMLS levels plus an explicit analytical join: nationally, visits fell and only partly recovered while e-circ rose; in most states those paths diverge; in seven, both indexes remain below the FY 2019 baseline.

For state librarians, mayors, and budget officers, the actionable cut is the quadrant — not a single national cheer or dirge. Where the substitution path holds, the capital question is how to fund buildings and digital licenses without pretending one replaced the other. Where dual decline holds, the question is harder: how to rebuild both the walk-in franchise and the digital one in places where neither bounced.