Trends

How the sector has grown — and where it’s headed

Key finding Nonprofit formation in the Inland Northwest is surging. After a pandemic dip in 2020 (179 new IRS rulings), new formations climbed to 310 in 2024 and 395 in 2025 — the strongest two-year run on record, even counting the one-time administrative spike of 2014, when the IRS introduced the streamlined Form 1023-EZ and processed a backlog of reinstatements. The 2010s produced more new nonprofits (about 2,030 IRS rulings) than any prior decade, and the 2020s are on pace to exceed them.

GU = ({
  navy: "#002967",
  red: "#c8102e",
  s1: "#1c5cab",
  s2: "#1baf7a",
  s3: "#c8102e",
  s4: "#eda100",
  ink: "#1a1a1a",
  muted: "#898781",
  grid: "#e1e0d9",
  seq: ["#cde2fb", "#9ec5f4", "#6da7ec", "#3987e5", "#256abf", "#184f95", "#0d366b"]
})

fnum = d3.format(",")

fmoney = (x) => x == null || isNaN(x) ? "—"
  : Math.abs(x) >= 1e9 ? "$" + d3.format(".2f")(x / 1e9) + "B"
  : Math.abs(x) >= 1e6 ? "$" + d3.format(".1f")(x / 1e6) + "M"
  : "$" + d3.format(",.0f")(x)

fpct = (x) => x == null || isNaN(x) ? "—" : d3.format(".1f")(x) + "%"

formations = FileAttachment("data/formations.csv").csv({typed: true})

orgs = FileAttachment("data/orgs_active.csv").csv({typed: true})

New nonprofit formations by year

Each point is the number of organizations receiving IRS tax-exempt rulings that year, across all seven counties — including organizations that have since dissolved. The red line is a five-year moving average. 2026 is a partial year and is excluded. The 2014 spike (467 rulings) is largely administrative: the IRS launched the short-form 1023-EZ application that July and worked through a backlog of reinstatements following the 2011–2012 automatic revocations.

formByYear = {
  const roll = d3.rollups(
    formations.filter(d => d.ruling_year >= 1960 && d.ruling_year <= 2025),
    v => d3.sum(v, d => d.n),
    d => d.ruling_year
  ).map(([year, n]) => ({year, n})).sort((a, b) => a.year - b.year);
  for (let i = 0; i < roll.length; i++) {
    const w = roll.slice(Math.max(0, i - 4), i + 1);
    roll[i].ma5 = d3.mean(w, d => d.n);
  }
  return roll;
}

Plot.plot({
  height: 380,
  width: 900,
  marginLeft: 50,
  marginBottom: 45,
  style: {fontSize: "13px"},
  x: {label: "Year of IRS ruling", labelAnchor: "center", tickFormat: "d"},
  y: {label: "New organizations", grid: true},
  marks: [
    Plot.lineY(formByYear, {x: "year", y: "n", stroke: GU.s1, strokeWidth: 1.5, curve: "monotone-x"}),
    Plot.dot(formByYear, {x: "year", y: "n", fill: GU.s1, r: 2}),
    Plot.lineY(formByYear, {x: "year", y: "ma5", stroke: GU.red, strokeWidth: 2.5, curve: "monotone-x"}),
    Plot.ruleY([0], {stroke: "#c3c2b7"}),
    Plot.tip(formByYear, Plot.pointerX({
      x: "year",
      y: "n",
      title: d => `${d.year}\n${fnum(d.n)} new rulings\n5-yr average: ${d3.format(".0f")(d.ma5)}`
    }))
  ]
})

md`<span style="color:${GU.muted}; font-size:0.85rem;"><span style="color:${GU.s1}">━</span> annual rulings &nbsp;&nbsp;<span style="color:${GU.red}">━</span> 5-year moving average. Hover for values.</span>`

Source: NCCS Business Master File, all vintages 1989–2026. Rulings before ~1989 appear only if the organization survived into the BMF era, so early decades are undercounted; the recent surge is not an artifact of this.

Formations by decade

byDecade = d3.rollups(
  formations.filter(d => d.ruling_year >= 1960 && d.ruling_year <= 2025),
  v => d3.sum(v, d => d.n),
  d => Math.floor(d.ruling_year / 10) * 10
).map(([decade, n]) => ({decade: `${decade}s`, n, partial: decade === 2020})).sort((a, b) => d3.ascending(a.decade, b.decade))

Plot.plot({
  height: 300,
  width: 900,
  marginLeft: 60,
  style: {fontSize: "13px"},
  x: {label: null},
  y: {label: "New organizations", grid: true},
  marks: [
    Plot.barY(byDecade, {
      x: "decade",
      y: "n",
      fill: d => d.partial ? GU.red : GU.s1,
      rx: 3,
      insetLeft: 6,
      insetRight: 6,
      tip: true,
      title: d => `${d.decade}${d.partial ? " (through 2025 — six years)" : ""}\n${fnum(d.n)} new rulings`
    }),
    Plot.textY(byDecade, {
      x: "decade",
      y: "n",
      text: d => fnum(d.n) + (d.partial ? "*" : ""),
      dy: -8,
      fill: GU.ink,
      fontSize: 12
    }),
    Plot.ruleY([0], {stroke: "#c3c2b7"})
  ]
})

* The 2020s bar (red) covers only 2020–2025 — six years — and has already passed three-quarters of the full 2010s total.

Where growth is happening

Spokane County still produces the most new nonprofits in absolute terms, but Kootenai County’s share of regional formations has risen with its population boom. The chart compares each county’s formations per five-year period.

countyGroups = formations
  .filter(d => d.ruling_year >= 1990 && d.ruling_year <= 2024)
  .map(d => ({
    ...d,
    group: ["Spokane County", "Kootenai County"].includes(d.county) ? d.county : "Other five counties",
    period: `${Math.floor(d.ruling_year / 5) * 5}–${String(Math.floor(d.ruling_year / 5) * 5 + 4).slice(2)}`
  }))

byPeriod = d3.rollups(
  countyGroups,
  v => d3.sum(v, d => d.n),
  d => d.period,
  d => d.group
).flatMap(([period, groups]) => groups.map(([group, n]) => ({period, group, n})))
  .sort((a, b) => d3.ascending(a.period, b.period))

Plot.plot({
  height: 360,
  width: 900,
  marginLeft: 55,
  style: {fontSize: "13px"},
  x: {label: null},
  y: {label: "New organizations per 5-year period", grid: true},
  color: {
    legend: true,
    domain: ["Spokane County", "Kootenai County", "Other five counties"],
    range: [GU.s1, GU.s3, GU.s2]
  },
  marks: [
    Plot.lineY(byPeriod, {
      x: "period",
      y: "n",
      stroke: "group",
      strokeWidth: 2.5,
      curve: "monotone-x"
    }),
    Plot.dot(byPeriod, {
      x: "period",
      y: "n",
      fill: "group",
      r: 4,
      stroke: "#ffffff",
      strokeWidth: 1,
      tip: true,
      title: d => `${d.group}\n${d.period}: ${fnum(d.n)} new rulings`
    }),
    Plot.ruleY([0], {stroke: "#c3c2b7"})
  ]
})

What kinds of nonprofits are being formed now?

Comparing the category mix of organizations formed since 2015 against the longer-standing sector shows where founders see unmet needs: recent cohorts tilt toward human services and away from the fraternal, membership, and religious organizations that dominate older cohorts.

cohortMix = {
  const active = orgs.filter(d => d.ruling_year != null && d.subsector_name !== "Unclassified");
  const mk = (rows, cohort) => {
    const total = rows.length;
    return d3.rollups(rows, v => v.length, d => d.subsector_name)
      .map(([subsector, n]) => ({subsector, cohort, share: n / total * 100, n}));
  };
  return [
    ...mk(active.filter(d => d.ruling_year < 2015), "Formed before 2015"),
    ...mk(active.filter(d => d.ruling_year >= 2015), "Formed 2015 or later")
  ];
}

Plot.plot({
  marginLeft: 200,
  marginBottom: 45,
  height: 430,
  width: 900,
  style: {fontSize: "13px"},
  x: {label: "Share of cohort (%)", labelAnchor: "center", grid: true},
  y: {label: null},
  color: {
    legend: true,
    domain: ["Formed before 2015", "Formed 2015 or later"],
    range: [GU.s2, GU.s1]
  },
  marks: [
    Plot.dot(cohortMix, {
      x: "share",
      y: "subsector",
      fill: "cohort",
      r: 6,
      stroke: "#ffffff",
      strokeWidth: 1.5,
      sort: {y: "-x"},
      tip: true,
      title: d => `${d.cohort}\n${d.subsector}: ${fpct(d.share)} (${fnum(d.n)} orgs)`
    }),
    Plot.ruleX([0], {stroke: "#c3c2b7"})
  ]
})

Active organizations with a classified NTEE category only. Source: NCCS BMF, June 2026 vintage.