Downtown Bellingham Business Health

1Scope and approach

The report tracks eighteen indicators in six groups, each with at least two years of data before and after the pandemic. Wherever two independent sources exist, both are used, so that no single file carries a conclusion.

Four kinds of count, never mixed. Jobs are county employers.³ Sales and business counts are Bellingham accounts reporting taxable sales.⁴ Licenses and locations are City registrations.⁵ The 2026 closure list is storefronts confirmed in news reports.

Definitions. Downtown is the City's Main Street Boundary, the 61-block area the City uses for its downtown business licensing data. Every business count, rate and map in this report uses it. The City's October 2025 vacancy survey used a different boundary of its own, which reaches further southwest and along the waterfront; the vacancy figures in Section 4 describe that area, and the report says so where they appear. A storefront is a business in retail, restaurants, bars and taprooms, personal services, or arts and recreation; upper-floor suites, apartments, food trucks and online-only sellers are removed. A location is a distinct address, counting each suite in a multi-tenant building separately, so a salon with several chair renters counts once. Real dollars are adjusted to 2025 prices with the Seattle-area consumer price index.⁶

What is downtown data and what is context. This is a downtown report, and only some public sources can be cut to the downtown boundary. The table says which, so that no citywide or countywide figure is mistaken for a downtown one.

How change is measured. Every indicator reports the same four numbers, the standard set used in economic indicator reports:

• Compared with the 2015–2019 average: percentage change from the typical pre-pandemic year. This is the main measure of where things stand against normal. An average is used rather than 2019 alone because a single year can be unusually high or low (2019 had the fewest openings of the period and the highest restaurant sales).
• From the year before: percentage change from the prior year, which shows the current direction.
• Pre-2020 range: the lowest and highest values in 2015–2019. A latest value outside that range is a real change; inside it is ordinary year-to-year variation.
• Trend per year: a straight line fitted through every year from 2015 (2016 for tax collections) to the latest, expressed as the yearly change as a share of the period's average, or in percentage points for rates. It describes the years since 2015, not two endpoints, so one unusual year moves it only a little. The summary table calls a trend up or down only when the slope is distinguishable from zero at the 95% level; otherwise it says “no clear trend,” even when the fitted line is not flat. Openings are an example: the fitted trend is slightly down (−2% a year) even though 2025 was the best year since 2015, because 2023 and 2024 were weak.

Rates, which are already percentages, change in percentage points (the closure rate going from 11.6% to 6.6% is −5.0 pts). City closure records use the 2015–2018 average, because the City changed how it records closures in late 2018;⁷ those comparisons are marked *. Local tax collections use 2016–2019, the first full years available. Real dollars are adjusted for inflation to 2025 prices. In the summary table, “trending up” or “down” is that fitted-line test alone. The written conclusions use a stricter standard: a change is called a trend only when the latest value is outside its range and a second, independent source agrees on direction.

2Business activity

2.1 Openings

Openings ran at 39–48 a year before the pandemic. In 2025 there were 47, the most since 2015: 7% above the 2015–2019 average and 38% more than in 2024. The dip in 2023 and 2024 has reversed. Openings are recorded promptly, so recent years are close to complete. In 2026, 30 opened through October 6, within the 21–38 range for the same months of 2020–2025.

2.2 Closure rate

Downtown storefront closures averaged 35 a year in 2015–2018 and 24 in 2020–2025. The rate has been below 7% since 2023 (6.3%, 5.7%, 6.6%), under the 2015–2018 range of 8–14%. Because pre-2019 counts may not be comparable, this rules out a large increase in closures but not a small one. Recent years may still edge up as late closures are filed, but the lag is bounded: 12 of the 13 closures reported in the news in 2024 and 2025 are now in the records (the thirteenth shares a license with an open sister restaurant), and the City's count of active restaurant licenses has tracked the state's count of restaurants filing sales tax within a few percent every year, including 2025 (Section 9).

Comparing downtown with the rest of the city is the stronger test, because the records' lag affects both equally. Downtown retail and personal services close less often than their counterparts elsewhere; downtown restaurants close more often (8.9% against 6.9% a year), a gap within the margin of chance.

Caveat. The closure rate is the one indicator in this report that depends on closures reaching the records, and the most recent year is always the least complete. Two checks bound the gap: 12 of the 13 closures reported in the news in 2024 and 2025 are now recorded (the thirteenth shares a license with an open sister restaurant), and the City's count of active restaurant licenses tracks the state's count of restaurants filing sales tax within a few percent (Section 9). That makes "closures are not up" a safe reading. "Closures are down" is provisional until the 2025 and 2026 figures are re-run in spring 2027.

3Downtown's place in the city

3.1 Storefront locations

Downtown had 271 storefront locations in January 2026, 7% above its 2015–2019 average and above that range of 238–261. The count fell in 2020 and recovered by 2023.

3.2 Downtown and the rest of the city

Downtown's share of Bellingham's storefront locations fell from 14.4% in 2016 to 11.5% in 2026, falling in nine of those ten years; the exception, 2023, was a rise of less than a tenth of a point. Downtown is not shrinking; the rest of the city is growing around it. This compares downtown with the rest of the city as a whole; no other district or urban village is ranked. Because the share falls whenever the rest of the city grows, whatever happens downtown, it is reported here as context and is not part of the summary table or the index. Part of the rest-of-city growth is home-based businesses, but restaurants, which need a commercial kitchen, show the same pattern: downtown restaurant locations went from 94 to 95 between 2019 and 2026 while the rest of the city went from 251 to 301.

3.3 Storefronts by sector

Retail has the longest decline: 91 downtown locations in January 2026 against a pre-pandemic average of 102, and below that average every year since 2020. Salons, barbers and spas are the one growing downtown group, from 47 to 61 locations since 2019 (licenses rose faster, from 78 to 128, because each chair renter holds one). Restaurants, bars and taprooms hold 95 locations, above their pre-pandemic range, though the rest of the city has added them faster since 2023. Jobs located downtown are in Section 3.4.

The rest of the city is above its pre-pandemic average in every sector. Part of that is home-based businesses, which register at a house, but the pattern holds for restaurants, which need a commercial kitchen.

3.4 Jobs located downtown

This is the one source that places jobs inside downtown itself.⁸ Downtown workplaces held 8,777 jobs in 2023, +16% against the 2015–2019 average and above that range, and 17% of all jobs in the city, up from 14% in 2015. Most of the growth is office, health and public employers, not storefronts.

Downtown's restaurants, bars and hotels employed 1,471 people in 2023, +7% against their pre-pandemic average and 26.2% of the city's restaurant and hotel jobs, a share that has been 24–27% in every year since 2016 except 2020, when it fell to 20%. Downtown retail employed 615 in 2023, −12% against its average: the job counts tell the same story as the location counts in Section 3.3, with retail the sector that shrank. These figures end in 2023, two years before the sales decline in Section 6 deepened, so they describe the base downtown entered 2024 with rather than where it stands today.

3.5 Which streets

Railroad Avenue lost restaurants (14 licenses to 10) and retail (15 to 12). State Street's salons consolidated into fewer addresses. Cornwall Avenue added restaurants and salons, and Commercial Street doubled. Closure rates by street are within chance of each other; the differences come from what opens, not what closes.

Three in four storefront openings and recorded closures since 2020 happened on Holly Street, State Street, Cornwall Avenue, Railroad Avenue or Commercial Street. Turnover is a sign of an active block as much as a struggling one: Holly and State have the most openings as well as the most closures. Railroad Avenue stands out this year, with four of the ten known 2026 closures (Avellino, Bellingham Antique Mall, Woods Coffee and Country Patty's), two of which already have a new tenant.

4Vacant space

4.1 Ground-floor vacancy

The City's field survey counted 193,239 square feet of vacant ground-floor space out of 1,935,582, a rate of 10%; 14.6% in the retail core.¹¹ The City's study cites industry guidance that treats 5–10% as healthy and above 11% as a warning sign. Only 44% of the vacant space is ordinary market-ready space; 19% is in large footprints and 37% in stalled properties, held up by ownership disputes, contamination or financing.

A different boundary. The survey used its own boundary, not the Main Street Boundary used for the business counts in this report: it reaches further southwest, to Laurel, Rose and Berry streets, and takes in the waterfront blocks along Granary Avenue. The difference matters less than it sounds: 42 of the 45 vacant spaces, holding 96% of the vacant square footage, are inside the Main Street Boundary too. The three outside it (two on N Forest Street, one on Granary Avenue) total about 6,900 square feet. The 10% vacancy rate still describes the survey area, because the City’s total of all ground-floor space is not published by block.

What the vacant spaces were. The City's space-by-space list, matched to the business records, shows that vacancy is not mainly a storefront-closure problem.¹²

• Three buildings are a quarter of it. The former JC Penney (20,625 sq ft), Rite Aid (13,718) and Greenhouse Home (10,800) hold 45,000 of the 193,000 vacant square feet, and all three are among the stalled or large-footprint properties the City names.
• A third of the spaces were offices or banks, which the storefront counts in this report never included. Vacant storefronts in the usual sense (former shops, restaurants and bars) are 23 of the 45 spaces.
• Most had been empty a long time. Of the 16 spaces whose previous tenant, as the City names it, can be found with a closing date in City records, 10 had been vacant for more than two years by October 2025. The recent closures (Ramble Tamble, Bayou on Bay, Time & Materials, Taco Lobo) were the exception, and two of those four already have a new tenant.
• A year on, 5 of the 45 have a new license at the address, three of them storefronts: Boundary Bay's taproom in the former Time & Materials space, Bandidos Cocina in the former Taco Lobo, and a nail salon on W Champion. Long-vacant space refills slowly; that is the pattern Section 4.2 measures for ordinary closures.
• The records can't see a dark window. In 16 of the 45 spaces the City found an active license at the address, usually another tenant in the same building or a business that had moved; in 5 the previous tenant the City named still holds one. That is why this report treats the walk as the source for vacancy and the records as the source for trends.

A second source gives the trend the City's single count cannot. Pacific Continental Realty, the brokerage that publishes Bellingham's quarterly commercial market report, put downtown retail vacancy at 8.4% in the third quarter of 2024, 9.5% at the end of 2024, 9.6% in early 2025, 10.3% in the third quarter of 2025 and 12.2% in early 2026, then 11.8% at mid-year and 12.1% in the third quarter of 2026, when citywide retail vacancy fell.¹³ The City's October 2025 count sits between those readings. Both sources say the same thing: downtown retail vacancy was at the top of the healthy range in late 2024 and has moved above it since, while citywide retail vacancy stayed near 5–6%. The City's records point the same way once the lag is allowed for: the share of downtown addresses that held a storefront license in the previous three years but have none today is 16%, about where it was before the pandemic (19–20% in 2016–2018) and down from 30% in 2020, but that figure cannot yet see this year's closures.

4.2 How fast closed storefronts refill

Among storefronts that closed in 2020–2022, 35% had a new business licensed at the same address within a year, and 37% for 2023–2025 closures. Spaces that refilled took a median of four to five months. Restaurant spaces refill fastest (47% within a year), retail slowest of the larger groups (31%); arts and recreation had too few closures to measure. The new businesses in those spaces tend to stay (Section 5.3).

5Business survival

5.1 How long new businesses last

Of storefront businesses that opened downtown in 2015–2019, 43% were open four years later, close to the rest of the city (46%). Restaurants were the exception: 39% downtown against 54% elsewhere, a gap at the edge of what chance would produce, though with 57 downtown restaurants in these years it is not a certainty (roughly a 1-in-19 chance of arising at random).

5.2 Who is closing

Closures of businesses under two years old fell from 46% of all closures in 2015 to 27% in 2025; closures of businesses ten or more years old rose from 15% to 23%. At least five of 2026's ten known downtown closures were ten or more years old: Temple Bar (26 years), Avellino (about 20), Woods Coffee on Railroad (since 2008), Cute But Crazy Socks (nearly 15) and Rock and Rye (since 2014).⁹

5.3 Do refills last?

A refilled space is good news only if the new business stays. Of the 49 downtown storefronts that opened in 2020–2024 in a space another business had left within the past year, 96% were still open a year later and 90% after two years, against 87% and 76% for the 135 other new downtown storefronts. Of the 52 storefronts that closed in 2020–2025 and refilled within a year, the new business closed within six months in 2 cases.¹⁴

New downtown storefronts as a whole last about as well as those elsewhere. In each opening year from 2020 to 2025, 82–100% were open a year later, against 81–93% in the rest of the city; 2024 was the one year downtown trailed (82% against 90%).

Caveat. Recent closures reach the records late, so these rates run high for 2024 and 2025 openings. The lag affects refills and other openings alike, so the comparison between them holds. They are not compared with 2015–2019, because closures may have been recorded differently before the 2019 jump in recorded closures.

6Restaurants and bars: citywide sales, county jobs, downtown closures

6.1 Restaurant sales

Restaurant sales peaked at $340 million in 2019 (2025 dollars), recovered to $334 million in 2023, and fell to $312 million in 2025, level with 2015: 4% below the 2015–2019 average and 4% below 2024. The first quarter of 2026 was 1.5% below the first quarter of 2025 before inflation.

Total sales are a volume measure, so a city that adds restaurants can hold its total up while each restaurant does less. That is what the second bar shows. Against their own pre-pandemic averages, sales per restaurant fell in nine of the eleven cities, by 7–21%; Bellingham's −11% is in the middle. What sets Bellingham apart is the first bar. It added restaurants slowly (8% more accounts reporting than before the pandemic, against 25–35% in Kennewick, Wenatchee, Mount Vernon, Ferndale and Lynden), so its total slipped 4% while cities that added restaurants faster grew. Of the seven other hubs, three saw totals fall, all further than Bellingham's (Olympia −13%, Everett −10%, Ellensburg −6%), and four grew (Vancouver +8%, Kennewick +5%, Mount Vernon +9%, Wenatchee +17%).

The recent direction is still Bellingham's own: it had the largest drop from 2024 to 2025 of any city here except Blaine (−3.8%), and it was the only one besides Blaine whose sales fell in the first quarter of 2026 (−1.5% against the first quarter of 2025, before inflation, while the others rose 1.5–8.3%).

6.2 Sales per restaurant

More restaurants are sharing less. The number reporting sales rose from 312 in 2019 to 329 in 2025 while real sales fell, so sales per restaurant dropped to $947K, below the pre-pandemic range of $1,033K–$1,089K. Sales per restaurant fell by a similar amount in nearly every comparison city (Exhibit 20), so more restaurants sharing less is a regional pattern, not a Bellingham one. Bellingham's total fell because it added restaurants more slowly than most of those cities, and because it is still losing sales in 2026. One definitional note: the state counts every account that reported restaurant sales at any point in the year, so a restaurant open for three months counts the same as one open all year. Years with more turnover therefore show slightly lower sales per restaurant for that reason alone; it does not change the direction.

6.3 Bar sales

Bar sales were already falling before the pandemic, from $29.4 million in 2016 to $25.8 million in 2019, and have held near $21 million since 2022, 25% below the 2015–2019 average.

6.4 Jobs per restaurant

Jobs per restaurant fell to 15.1 in 2025, 8% below the 2015–2019 average and below the pre-pandemic range of 16.2–16.7. That is a regional pattern, not a local one: peer Washington counties fell by a median of 7.5%, and Kootenai County, Idaho, by 17.5%. Full-service restaurants in Whatcom fell less than in any peer county (−4%, against −5% to −26%; peer median −14%). Total restaurant employment in Whatcom is down 5% since 2019 while peers are up 2%, which fits the sales decline.

6.5 Downtown restaurant and bar closures

City records show 5–11 downtown restaurant and bar closures a year in 2020–2025. At least seven have closed in 2026 through October 6 (eight counting Welcome Road's wine and coffee bar). With three months left and records still catching up, 2026 is on pace to finish near the top of that range.

Caveats. Job figures are countywide; sales figures are citywide. Jobs count employees, not hours or owners. The inflation adjustment uses all-items prices; menu prices rose faster, so the drop in meals served is likely larger than the dollar figures show. Restaurant pay per worker rose 46% from 2019 to 2025, about 12% after inflation.³

7Context: visitors and the wider economy

7.1 All taxable sales

Local sales tax from all businesses is within its 2016–2019 range after inflation, 3% above the 2016–2019 average but 2% below 2024.¹⁵ Collections rose 6% in the first nine months of 2026, but part of that is a tax change: since October 2025 the state taxes several services that were not taxed before, and Ferndale, under the same change, rose 1%.¹⁶ Ferndale's collections grew 16% after inflation from 2019 to 2025.

7.2 Hotel and motel tax

Hotel tax recovered above its pre-pandemic range in 2022 and 2023, then fell two years in a row to $1.11 million, 5% below the 2016–2019 average and 8% below 2024. Collections in the first nine months of 2026 were 1% above the same months of 2025.

7.3 Border crossings

Crossings were 5.2–5.8 million a year (January–August) before the pandemic. They recovered to 4.4 million in 2024, fell to 3.2 million in 2025, and stand at 3.4 million in 2026.¹⁷

82026 so far

At least 10 downtown storefronts have closed this year and five of those spaces already have a new business. Openings are normal. Closures have not reached the records yet, but known restaurant and bar closures are already within the range of a full year.

8.1 Year to date, all measures

Closure counts for earlier years come from City records, which date about a quarter of closures December 31; 2026 counts come from news reports. The two are not exactly comparable, which is why the restaurant comparison uses full-year totals as well as the January–October window.

Compiled from Cascadia Daily News, My Bellingham Now, The Bellingham Herald and Salish Current, matched to City and state records. Not listed: Red Star Taco Bar (the concept closed September 25; Sweet Bay Cafe stays open in the same space) and Mount Bakery downtown (announced, not yet closed). Boundary Bay Brewery's Railroad Avenue brewpub is a 2025 closure in City records (December 31, 2025); its State Street taproom, licensed January 1, 2026, counts as a 2026 opening. Some news coverage cites larger 2026 counts; this list includes only downtown storefronts whose closure is confirmed in news reports and matched to City or state records, and leaves out announced closures and concepts that closed inside a business that stays open.

9How far the records can be trusted

The City's registry is accurate about openings and current businesses, but records closures late and misses moves. Three checks measure how far off it is.

City licenses against state tax filers, year by year

The third check in full. The City column counts restaurant licenses (industry code 7225) active on July 1 of each year, citywide. The state column counts Bellingham restaurants that reported taxable sales to the Department of Revenue that year. A ratio above 1.00 would mean the City is carrying licenses for restaurants that no longer sell anything, which is what a backlog of unrecorded closures looks like.

The ratio has sat just under 1.00 for a decade; 2025 (0.98) matches 2018 and 2021. The City count runs slightly below the state's because a few restaurants file sales tax under a different industry code than the one on their City license. Source: City of Bellingham business registry; Washington Department of Revenue, Taxable Retail Sales, Bellingham, NAICS 7225, accounts reporting.

Why restaurants only. The check needs two counts of the same businesses, and restaurants are the one storefront sector where the state's count is a count of local establishments. For retail, the state's Bellingham count rose from 3,930 accounts in 2015 to 12,714 in 2025 while City storefront retail licenses stayed between 879 and 1,019: since 2018, out-of-state online sellers report sales by the city they ship to, so the state count is mostly remote sellers. For personal services, haircuts and similar services are not subject to sales tax, so most salons never appear in the state's count and the City count runs about two-thirds above the state's every year. Arts and recreation has the same problem in a milder form. Those sectors rely on the other two checks above.

The earliest years are thin. City records show only 23 downtown storefront closures in 2015 and 4 restaurant closures, far below later years. The 2015–2018 comparison for closures therefore leans on 2016–2018, and the pre-2020 range for restaurant closures (4–12) is wider than it would be with complete records.

Why the records lag, and how long. Opening a business requires registering, so openings appear right away. Closing requires nothing; a record closes only when the owner tells the state, often months later. The news check above puts a number on it: closures reported more than a year ago are all in the records, and closures reported this year are mostly not yet. The lag is real and it is bounded at about a year, so the last twelve months of closures are always undercounted and earlier years are close to complete. Close dates are often paperwork dates: about a quarter of downtown storefront closures since 2015 are dated December 31.

Duplicate registrations. Some businesses hold two overlapping licenses at the same address. These are merged and counted once throughout this report (for example, Boundary Bay's two Railroad Avenue licenses are one 2025 closure).

10Open questions and next steps

Indicators to track. Future updates should report the eighteen indicators in the summary table against the same pre-pandemic ranges, add ground-floor vacancy once a second count exists, and treat everything else as description. This avoids finding a trend by chance among dozens of series.