Every Number in the Green

Year Twelve, September. Missing Person Report 24-0917. Status: Open.

The coffee in the mug on the desk had gone cold, then room temperature, then developed a faint skin across the surface that the responding officer noted in his report as “coffee, cold, approximately two days.” The municipal building’s fluorescent lights cycled at sixty hertz and nobody in the office could say whether they had always made that sound or whether it had started when Loreen Voss stopped coming to work.

She had not come to work on Tuesday. She had not called. Her car, a 2019 Subaru Outback with 43,000 miles and a Garner Falls Volunteer Fire Department sticker on the rear window, was parked in its usual spot in the municipal lot, space 7, backed in as she always backed in. The doors were locked. A reusable grocery bag from the Saturday farmers’ market sat on the back seat. A parking pass hung from the rearview mirror, valid through December.

Her office was orderly. The filing cabinet was closed. The quarterly report for July through September had been submitted on time and was, by every measure the town council would later apply, accurate. Crime was down 38 percent from her first year. Test scores at Garner Falls Middle School held at the 94th percentile statewide. The water quality index had not produced a single advisory in twelve consecutive years. Property values had risen 3.1 percent annually for the past nine, a figure so consistent that the county assessor’s office had once called to ask if it was a projection. It was not a projection.

Her personal laptop was open on the desk. The screen had gone to sleep, and when Officer Denise Paulson touched the trackpad, it showed a satellite image of Garner Falls with a series of lines drawn in red. Seven lines, not connecting to each other, running through yards and across vacant lots and into the strip of trees behind the middle school. Paulson noted this in her report. She described them as “red markings on a map, purpose unclear.”

Janet Caulfield, who worked in the permits office across the hall, told Officer Paulson that Loreen had seemed fine. Happy, even. She’d brought scones on Tuesday of the previous week — from Bev’s bakery, cranberry-orange, the kind she always brought — and mentioned she was thinking about a new bookshelf for her living room. She had been taking walks in the evenings lately, Janet said, though she didn’t know where. Odd hours. After dinner, sometimes after dark. Janet had seen her once from her own kitchen window, crossing the vacant lot on Birch Street at a quarter to ten, walking without a flashlight in a direction that did not lead to any street or building Janet could think of.

“She stopped wearing her watch,” Janet said. “Maybe a week before. I asked her about it and she said she was trying something.”

“Trying what?” Paulson asked.

“She didn’t say. I didn’t ask.”

Paulson noted this. She also noted the state of the desk: clean, organized, a pen in the cup, a coaster under the mug, a framed photo of the Garner Falls river trail in autumn. No personal effects missing. No drawers emptied. No sign that Loreen had been planning to leave. The desk of a person who intended to come back.

The investigation would be thorough. Garner Falls had the resources for thorough — a department of eleven officers, well-funded, with a solve rate that Loreen herself had calculated at 94 percent for non-traffic incidents. Missing persons cases, Loreen’s own data showed, were rare. This was the first in six years. The last one had resolved in forty-eight hours: a teenager at a friend’s house in Bend. The case before that, in 2016, had been a hiker on the ridge trail above town. Found within the day, dehydrated, grateful. Garner Falls did not lose people.

The municipal building was quiet that afternoon. Someone from Public Works replaced the fluorescent tube in the hallway that had been flickering since Monday. The new tube did not flicker. Outside, the elms along Main Street were beginning to turn, and the volunteer fire department was putting up banners for the Fall Festival, and the parking meters had been recalibrated the previous spring to give an extra fifteen minutes per quarter, a policy Loreen had recommended based on turnover data.

The red lines on her laptop screen did not lead anywhere the satellite could show.

Year Ten. Annual Municipal Report: All Indicators Positive. Eighth Consecutive Year.

The town council met on the second Thursday of October, and Loreen presented her annual report from the same podium she had used for the past nine presentations, a wooden lectern with a brass plaque that read IN SERVICE TO THE COMMUNITY OF GARNER FALLS. The council members sat at a curved table with name plates and water glasses, and the public seating held eleven residents, which was a typical number for a meeting without a contentious agenda item.

She had prepared thirty-two slides. Property values, crime statistics, school performance, water quality, business permits filed, voter turnout, fire inspections completed. Each metric was positive or stable. She paused on the slide showing the ten-year trend for voter turnout — 74 percent, against a national midterm average of 42 — and Councilwoman Deborah Hines said, “That’s something to be proud of,” and the room murmured agreement.

Loreen felt the satisfaction she usually felt when the presentation went well, which it always did, because the data always supported it. She also felt a tightness in her jaw, on the left side, where her molars met, that had started six months ago and had not responded to the night guard her dentist had prescribed. She felt a heat in both palms that she associated with the spreadsheets specifically — a rash that appeared when she opened the files and faded by midmorning, prickling and dry, as though the skin were trying to shed.

After the council adjourned, she sat in her office and ran Benford’s Law analysis on the town’s financial data. She had learned the technique in graduate school: in naturally occurring datasets, the leading digit “1” appears approximately 30 percent of the time, following a logarithmic distribution. Datasets that deviated from this pattern were flagged as potentially fraudulent. Tax authorities and forensic accountants used it as a first-pass detection tool.

The results for Garner Falls were textbook. Not merely compliant — textbook. The distribution of leading digits across twelve years of budget line items, revenue streams, and expenditure reports matched the expected Benford curve to four decimal places. In a dataset this small — a town of 8,400 — that degree of conformity was itself anomalous. Natural data was messy. Natural data showed noise, clusters, gaps. Data that perfectly matched a statistical law was, by the logic of that law, the least natural data possible.

Loreen noted this in her working file. She typed: Benford compliance: exact. Flag for review. She saved the file. She closed it. She heard the click of the laptop closing and then the quiet of the office around the click, and then the hum of the fluorescent lights, and then the building settling, and then nothing.

She would never open the file again. The flag for review would remain unflagged: a single line in a working document that no one else would read, noting that the town’s financial data was statistically impossible, filed between a note about park usage rates and a reminder to update the library circulation template.

Driving home on Winslow Road, she passed Prospect Park. A worn trail cut diagonally across the planned walking path, bare earth through the grass, wide enough that it had been walked by many feet over many months. She noticed it with the part of her mind that sorted patterns automatically, filed it, and did not slow down.

Her jaw ached. She pressed her tongue against her molars and counted the seconds until the pressure felt like relief. She counted to forty.

Year Eight. Property Values: +3.1%. Eighth Consecutive Year.

Paul Renner worked at the county assessor’s office in Salem and drove to Garner Falls on weekends. He had a truck with a toolbox bolted to the bed and opinions about property tax methodology that Loreen found both tedious and comforting. They had been seeing each other since April. He slept on the left side of the bed and kept a toothbrush in her bathroom and had never suggested he move in, which she appreciated with a precision she could have quantified if anyone had asked.

“Your town is weird,” he said one Saturday, not unkindly. They were at the kitchen table. Rain on the windows. She was making notes on the quarterly report. “Not the people. The values. They go up the same amount every year. Three point one percent. It’s like — ” He paused. “Like someone ironing a shirt. Same pressure, same heat, every pass.”

She laughed. “Stable growth.”

“Stable growth has variance. This is a metronome.”

She made a note to look at the comparable data from similar-sized Oregon towns. She would find that La Grande fluctuated between negative 2 and positive 6. That Hermiston spiked and corrected. That Garner Falls’ steadiness was, in the statistical sense, extraordinary. She would find this, and she would file it, and in the filing it would become a fact about Garner Falls rather than a question about it.

Paul mentioned, also, that the fire inspector — Ed Kessler — shared a last name with the man who owned the building where her office was. She said she’d noticed. They were cousins. She had checked. She did not say that she had checked only far enough to find the reasonable explanation, and no further.

That Saturday, she walked the desire paths. She had begun to track them the previous year, marking them on a satellite printout with a red pen. Seven paths, cutting through yards, across the vacant lot on Birch, along the drainage ditch behind the Safeway, through the strip of scrubby firs behind the middle school. They did not connect to each other. They did not form a route between any two useful points. They were not shortcuts. People walked them anyway — she could see the wear, the packed earth, the grass thinned to nothing in the centers and growing back at the margins.

She walked three of them end to end. The longest was perhaps four hundred yards and terminated at a chain-link fence behind a house on Alderman Street. Her street. She stood at the fence and looked at the back of a house she passed every day. The porch light was on. A child’s bicycle leaned against the railing.

She counted the footprints in a muddy section near the school. Thirty-eight distinct treads in both directions. She photographed the section and measured the width of the path at its broadest point: fourteen inches. Packed hard, the soil compressed in a way that suggested years of use, though when she checked the satellite history, the path behind the school appeared only in images from the past three years. Before that, unbroken grass. She recorded this. She did not ask any of her neighbors whether they walked the paths. She did not ask where the paths went or when they had started. The paths went where they went, and the wear was the data, and the data was sufficient.

Paul stopped coming to Garner Falls in November. He did not break up with her so much as stop making the drive, and she did not call to ask why, and this too became a data point she filed without analysis.

Her jaw was tight but not painful. Her palms were clear. She slept six hours most nights, waking once around 3 a.m. for a reason she cataloged as “no reason.”

Year Six. School Performance: 94th Percentile. State Average: 71st.

The conference was in Portland, at a hotel on the river with windows that ran floor to ceiling. Loreen presented her paper — “Sustained Positive Deviation: Twelve Indicators of Municipal Health in a Small Oregon Community” — to a room of thirty regional planners, city managers, and two graduate students taking notes.

The data was clean. The story it told was simple: Garner Falls had improved across every measurable dimension, consistently, for six consecutive years. Crime, schools, property values, voter engagement, infrastructure maintenance, water quality, business permits, library circulation, park usage, fire response times, traffic incidents, volunteer hours. Twelve indicators, all positive, all trending in the same direction.

A man in the third row — wire glasses, county lanyard — raised his hand during the Q&A.

“What explains it?”

“The data doesn’t tell me why,” Loreen said. “Only that.”

“Doesn’t the absence of explanation concern you?”

“No,” she said. “I measure outcomes. Outcomes are the domain.”

The man nodded. He did not press further. Someone asked about her methodology, and the conversation moved to sample sizes and confidence intervals, and the moment passed the way such moments always passed in Loreen’s professional life — absorbed into procedure, translated into a question about method rather than meaning.

That night in the hotel room she could not sleep. She was settled, productive, confident. She cooked elaborate meals in her kitchen on Alderman Street. She ran three miles along the river trail every morning. By every metric she would use to evaluate the well-being of a person in a given municipal environment — access to services, safety, social engagement, physical health, financial stability — she was well.

She opened her laptop at two in the morning and looked at the satellite image of Garner Falls. She had added four new desire paths since she started tracking. Eleven now. The new ones were faint, barely worn, but distinct from the satellite altitude. She zoomed in on the one behind the middle school. It curved in a way that did not follow the contour of the land. She moved to the one along the drainage ditch. Also curved. Not the curve a person takes when cutting through a yard — the body’s efficient line between two points. A different curve. Deliberate, or else following some feature of the ground she couldn’t see from sixty miles of altitude.

She counted them. Eleven. She toggled the image to a view from the previous spring. Nine. Two new paths in eight months.

She closed the laptop and lay in the hotel bed and listened to the river, which was not the same river that ran through Garner Falls but sounded the same, and she slept, eventually, with her jaw loose and her palms smooth and her body at ease in the clean sheets.

Year Four. Voter Turnout: 74%. National Average: 42%.

Bev Alderman brought scones to the municipal building every Tuesday. She had done this since before Loreen arrived, but Loreen had been the first person to keep count: 208 Tuesdays in four years, never missed. Bev ran a bakery on Main Street, the street her family’s name was on, and she had the particular confidence of a woman whose great-grandparents had built the building she worked in.

“I remember when this town was different,” Bev said. They were in the break room. The scones were cranberry-orange. The coffee was from the pot Loreen maintained, 12 cups brewed at 7:15 each morning. “Main Street had vacancies. The school almost closed — that would have been, what, 2002? The church was losing members. The usual slow drain.”

“What changed?”

Bev thought. She broke a scone in half and considered the question with the seriousness of someone who had been asked it before and had never arrived at a satisfying answer.

“It just got better,” she said. “Not all at once. Just every year, a little better.”

“When?”

“About the time we started measuring,” Bev said, and laughed. She picked up the other half of her scone. “I don’t mean it like that. I just mean it all seemed to happen at once. The school got better. The downtown filled up. People started volunteering. Like someone turned a dial.”

Loreen laughed too. The coincidence was just that — a coincidence, or else a natural consequence of attention. Towns that tracked their metrics tended to improve on those metrics. The Hawthorne effect, the observer effect, the principle that what gets measured gets managed. There was nothing unusual about improvement following measurement. That was the point of measurement.

She did not distinguish between these explanations. She did not need to. The data was the data. And Bev’s scones were good, and the break room had been repainted last spring in a color the facilities manager called “warm cream,” and the coffee was hot, and the morning light came through the window at an angle that made the linoleum floor look almost beautiful.

Her body was soft that year. Rested. She did not know what a tight jaw felt like, had never ground her teeth, had no rash on her palms. She slept eight hours and woke to the sound of the river through her open window and made coffee and walked to work along Alderman Street, past the houses with their kept lawns and their functioning gutters and their porch lights that came on at dusk by timer or by habit.

She had not yet found the desire paths, or they did not yet exist. The strip of grass behind the library was unbroken. The lot on Birch was tangled with blackberry canes, unwalked. She took the planned routes — the sidewalks, the crosswalks, the paved trail along the river — and she arrived where she intended to go, and the town smelled like pine and bread and cold water from the falls.

Year Two. Crime Rate: -12% Year-Over-Year.

She ran the numbers three times. Not because she suspected error but because the numbers were low in the way that makes a person recount, the way a cashier counts a drawer that comes up over and finds it correct and counts again anyway. Down 12 percent from the previous year, which had itself been down 8 percent from the year before that. Petty theft, vandalism, DUI, domestic disturbance — every category declining. Not plummeting, not impossibly so, just declining at a rate that was, in the vocabulary she had used at her previous job, “statistically noteworthy.”

At her previous job she had processed crime data for a county of 600,000 people. Numbers at that scale had texture. Spikes in summer, drops in January, neighborhoods that diverged from the trend, outliers that told stories the averages hid. Garner Falls’ numbers had no texture. They moved in one direction. They moved smoothly.

She drove the streets at night that winter, for no reason she could have articulated in the language of her profession. Past the hardware store, dark and locked, its window display of snow shovels unchanged since November. Past the library with the copper roof, green now, a patina that took decades. Past the diner where the same waitress worked Thursday through Sunday and remembered what Loreen ordered. Past houses with porch lights on and living room windows throwing warm parallelograms onto lawns where children’s bicycles lay on their sides, unchained, trusting the dark.

She felt something she filed under “adjustment.” The sensation of having arrived somewhere that was exactly what it claimed to be. She had spent four years in Portland processing data that was, by every measure, bad. Rising assaults. Declining clearance rates. Response times stretching. She had come to Garner Falls because the job listing described a town that was getting better, and she wanted to be in a room where the news was good.

The news was good. The news was always good. And she could not say whether the feeling in her chest was gratitude or the specific discomfort of having her deepest professional wish granted without conditions.

That spring, she found her first desire path. A worn line through the grass behind the library, narrow, the dirt packed smooth, curving away from the parking lot and toward the river. She followed it thirty yards to where it entered a stand of willows, the branches dragging, the light going green and close. The path continued into the trees, visible as a gap in the undergrowth. She stopped. She listened to the river. A bird she couldn’t identify called twice and didn’t call again. The willows moved in a way that was just wind.

She went back to her car. She did not think about the path again for months. When she did, it was because she noticed a second one — a faint track through the grass on the slope behind the elementary school, running parallel to the paved walkway but twenty feet uphill, where no walkway was needed. She noted it the way she noted everything: as information, as a fact about the town, as a line that might someday belong to a dataset she hadn’t built yet.

Her body was loose and new and unmarked. She slept nine hours. She had not yet begun to grind her teeth. Her palms were dry, in the ordinary way of winter, and she rubbed lotion into them each morning at her desk, a ritual as automatic and as comforting as the coffee she brewed at 7:15, and the data she opened at 7:30, and the green numbers that met her every time she opened the file.

Year One, March. Position Accepted. Start Date: April 1.

She drove into Garner Falls on a Thursday in late March, the mountains still carrying snow on their upper ridges, the river running fast and gray with melt. Route 26 dropped through forest and then opened, the way Oregon valleys open — suddenly, a widening of light, the trees pulling back to reveal a town that had been there the whole time, waiting below the timber line.

A hardware store with a hand-painted sign: KESSLER’S, since 1971. A diner called Rosie’s with a specials board facing the street. The library, set back from Main behind a row of Douglas firs, its copper roof turned green, its front steps swept. A church with a white steeple. A bakery with the door propped open, the warm yeast smell reaching the sidewalk. The falls themselves — not large, not dramatic, just steady, the river dropping fifteen feet over a basalt shelf at the south end of town, the sound always present, a white noise the town lived inside of.

She parked in front of the municipal building, a two-story brick structure with tall windows and a flagpole and a fire inspection certificate in a plastic sleeve taped to the glass door. The certificate was signed and dated. Current. Beside it, a flyer for a pancake breakfast to benefit the volunteer fire department, scheduled for the first Saturday in April. Below that, a water quality report from the Oregon Health Authority, printed on white paper and posted where anyone walking in could see it, every metric within the acceptable range, each number highlighted in green by someone who had found this reassuring enough to color-code.

The air was cold and smelled like snowmelt and pine resin and, faintly, bread. She stood on the sidewalk and looked at the town in both directions. A woman came out of the bakery with a white paper bag and said good morning. A pickup truck with a Garner Falls Volunteer Fire Department sticker on the bumper pulled into a space in front of the hardware store. The specials board at Rosie’s advertised meatloaf.

She had left Portland on a Tuesday. Packed her apartment in nine boxes and a suitcase, terminated her lease, filed the change-of-address form, forwarded her mail. Four years of processing crime data for a county that was not getting better. Four years of spreadsheets that trended the wrong way, quarterly reports that said the same thing in different numbers, meetings where nobody clapped. She had seen the Garner Falls listing on a state employment board — Senior Data Analyst, Municipal Planning, population 8,400, competitive salary, benefits — and she had applied because the job description mentioned positive indicators. Because the phrase “sustained improvement” appeared in the second paragraph. Because she was tired of measuring decline and wanted to measure something that worked.

The building was warm inside. The hallway had linoleum floors and bulletin boards with notices for recycling schedules and park hours and a meeting of the town council on the second Thursday of the month. Her office was on the second floor, facing Main Street, with a window that opened and a desk and a filing cabinet and a computer that was already set up. Someone had left a card on the keyboard: WELCOME TO GARNER FALLS, LOREEN! — signed by eight people, names she did not yet know.

She sat in the chair. It was a good chair — ergonomic, adjustable, newer than the desk it faced. She opened the computer. The desktop had a folder labeled ANNUAL REPORTS and she opened it and found twelve years of municipal data, neatly organized, each file named by year and category. She opened the most recent. Revenue. Expenditures. Permits. School enrollment. Crime. Water. Fire. Parks. Each metric graphed over time. Each graph rising, or falling in the ways that meant rising. The lines were clean. She scrolled through them slowly, the way a person reads a letter from someone they want to trust — looking for the catch, finding only good news, and the absence of a catch feeling, for the moment, like the best news of all.

Outside, the snow was melting. The river was running. The falls were audible through the open window. She opened the next file. The numbers were good.