Why Your SaaS Can Owe State Tax on Sales You Never "Made" Anywhere: The Throwback and Throwout Trap

You incorporated in one state, sell SaaS subscriptions to customers in 30 others, and have no office or employees outside your home base. You filed income tax only where you have nexus and assumed the rest was someone else's problem. Then your home-state return comes back with a question that doubles your taxable income: why didn't you "throw back" those out-of-state sales?

If you sell software, services, or anything delivered remotely, you can owe corporate income tax on so-called "nowhere income" — sales that no destination state has the right to tax, but that your home state happily claims as its own. The mechanism is a throwback or throwout rule, and for SaaS and multistate service businesses it is one of the least understood reasons a tax bill spikes after a good growth year.

This guide explains what nowhere income is, how throwback and throwout rules work in plain math, which states still use them, why SaaS founders are not protected by the federal shield most people cite, and what to track so you can model and manage exposure before an audit does it for you.

Apportionment in 60 Seconds: How States Decide What Share Is Theirs​

When you do business in more than one state, no single state gets to tax 100% of your profit. Each state taxes only an apportioned slice.

Historically, states used an evenly weighted three-factor formula: property, payroll, and sales in the state divided by property, payroll, and sales everywhere. If 20% of your property, 10% of your payroll, and 30% of your sales were in State A, State A taxed roughly 20% of your income.

That world is mostly gone. Today, 30-plus states use single sales factor — sales alone determine your apportionment — or give sales a heavy weight. The policy goal is explicit: reward companies that employ and invest locally but sell nationally, and tax companies that sell locally from elsewhere.

The math matters because every throwback and throwout adjustment targets the sales factor. Whether that factor is the whole formula or one-third of it, inflating the numerator or shrinking the denominator inflates the tax you owe in the throwback state.

For your books, this means the most important state-tax data you keep is not just revenue. It is revenue by destination state, consistently sourced, reconciled to your payment processor and your general ledger every month.

What "Nowhere Income" Means and Why It Exists​

A state can only tax you if you have nexus — a sufficient connection. Nexus comes from statute, the U.S. Constitution, and one critical federal limit: Public Law 86-272.

P.L. 86-272 says a state cannot impose net income tax on your sale of tangible personal property if your only activity in that state is soliciting sales and you ship via common carrier. No warehouse, no employee visiting, no owned truck delivering — just remote solicitation. In that narrow case, the destination state lacks jurisdiction even if you sell millions there.

The result is "nowhere income": profit attributable to a sale that the destination state cannot tax. Classic example: a manufacturer in Wisconsin with no property or payroll in Iowa ships widgets to an Iowa customer via FedEx and has no Iowa office or rep. Iowa cannot tax that income under P.L. 86-272, so the income is "nowhere" for apportionment purposes.

States that want 100% of corporate income taxed somewhere see that as a loophole. Throwback and throwout rules are their fix. For tangible goods, about 20 states plus the District of Columbia still throw those nowhere sales back into the origin state's numerator. Three states use the milder throwout variant. Five states — Alabama, Louisiana, Missouri, Vermont, and West Virginia — repealed their rules since 2019, and Arkansas began phasing its throwback out in 2024. The list changes every year, so you must verify the current roster when you file, not when you read this.

Here is the catch for SaaS founders: P.L. 86-272 does not protect sales of services, SaaS, or other intangibles at all. It was written for boxes moving on trucks. If you sell subscriptions, the destination state's ability to tax you depends on its own nexus thresholds — often a bright-line $100,000 of sales or 200 transactions after Wayfair, or a $250,000 receipts test in states like Arkansas — and on market-based sourcing rules, not on P.L. 86-272. That means you can easily create nowhere income with services without ever touching tangible-property rules, and you cannot rely on the one federal shield most blog posts tell you about.

Throwback vs. Throwout: The Same Goal, Very Different Math​

Both rules increase the home-state sales factor, but they do it in different places in the fraction:

Sales factor = In-state sales ÷ Everywhere sales

• Throwback: Nowhere sales are added to the numerator. If you are based in State A and have $1M of nowhere sales into states where you lack nexus, State A makes you add that $1M to State A sales for the factor.

• Throwout: Nowhere sales are removed from the denominator. State A leaves its numerator alone but subtracts that $1M from total sales.

Both raise the fraction; throwback raises it far more.

A SaaS Example With Real Numbers​

Assume you are a C-corp domiciled in a single-sales-factor state with a throwback rule. You have $5M in total sales:

• $1M to customers in your home state
• $2M to customers in states where you have nexus (you file there)
• $2M to customers in 12 states where you have no nexus and the destination state cannot or does not tax you

Without any rule, your home-state sales factor is:

$1M ÷ $5M = 20% → home state taxes 20% of your apportionable income.

With a throwback rule, the $2M of nowhere sales is thrown back:

($1M + $2M) ÷ $5M = $3M ÷ $5M = 60% → home state taxes 60% of your income, triple the starting point, even though you made no additional sales at home.

With a throwout rule, the $2M is excluded from the denominator:

$1M ÷ ($5M − $2M) = $1M ÷ $3M = 33.3% → still a 66% increase, but less than throwback.

If your home state instead uses three-factor apportionment with sales weighted one-third, the impact is diluted but still material — a 20-point swing in the sales factor becomes a 6- to 7-point swing in total apportionment, which on $1M of profit at a 6% rate is $3,600 to $4,200 you did not budget.

The same logic applies in reverse if you are the out-of-state seller: a throwback state where you do have nexus can pull in your nowhere sales from other states, raising your factor there too.

Which States Still Use These Rules — and Why SaaS Can't Ignore Them​

The 2024 map tells a clear story: throwback states remain the majority among those that have any rule, but the trend is away from them. States recently repealing or narrowing include Alabama, Louisiana, Missouri, Vermont, West Virginia, and Arkansas phasing out. That leaves roughly 20 states plus D.C. with a throwback for tangible property, and a handful with throwout.

For tangible goods, the list historically includes states such as California, Illinois, Massachusetts, Michigan, New Mexico, and Wisconsin — but you must check current statute because sourcing definitions shift.

For SaaS and services, most throwback statutes on the books still apply by their terms only to sales of tangible personal property. In those states, a pure SaaS subscription is typically sourced by market-based rules (where the customer receives the benefit) or, in a minority, by cost-of-performance (where the work was done), not by throwback. That does not make you safe. Three risks remain:

1. You sell more than SaaS. If you sell hardware, merch, printed materials, or any tangible item alongside subscriptions, the tangible portion is subject to throwback in those states.

2. Market-based sourcing can create its own nowhere problem. If you apportion based on where your customer is, and you have not established nexus there, your home state may still want that receipt somewhere — and combined-reporting rules like Joyce versus Finnigan determine whether affiliates' sales count.

3. Nexus for services is broader. Because P.L. 86-272 does not apply, you may have nexus in more states than you think via economic thresholds, making fewer sales truly "nowhere" but more states able to claim apportionment, which can push you into double-weighted nowhere recapture plus multi-state filing obligations.

Practical takeaway: do not ask "does my home state have a throwback rule?" Ask "does my home state have a throwback or throwout rule for each class of receipts I have, and how does it source my SaaS receipts in the first place?"

How Modern Service Sourcing Actually Works​

For services and intangibles, states choose between two philosophies:

• Market-based sourcing: the sale is sourced to where the customer receives the benefit — typically the customer's billing address or where the software is used. California, Georgia, New York, and most newer adopters use this. Under market-based, a sale to a customer in State B counts as a State B sale, even if your engineers are in State A. This is intuitive for SaaS but creates a trap: if you lack nexus in State B, that sale becomes nowhere income that a throwback state can reclaim.

• Cost-of-performance / income-producing activity: the sale is sourced to where you performed the work. A minority of states still use this. There, your SaaS sale might count as a home-state sale from the start, which ironically avoids nowhere treatment but piles income into your domicile regardless of where customers sit.

Some states blend these or have industry-specific rules. Add single-sales-factor elections and you can have the same dollar of SaaS revenue sourced to two states at once under conflicting rules — the mirror image of nowhere income, taxed twice. That is why states fight over 100% taxability: one state's nowhere is another state's double-count.

For combined groups, it matters whether the state uses Joyce or Finnigan rules. Under Joyce, each entity's nexus is tested separately; under Finnigan, the whole combined group is treated as one taxpayer. In a throwback state, Finnigan pulls more sales into the throwback net. As of recent data, among combined reporting states with throwback or throwout for tangible property, about 12 use Joyce and 8 use Finnigan, plus D.C. on Joyce. If you operate with multiple LLCs or a holding company, the election changes your exposure materially.

How to Model Your Own Exposure This Quarter​

You do not need a SALT firm to get a first-pass estimate. You need clean data and a simple model you rerun quarterly.

1. Build a sales-by-state ledger you actually trust​

Create a single source of truth — not a spreadsheet you update at year-end. For every invoice, subscription charge, or payout, capture:

• Customer ship-to / bill-to state (use the sourcing rule your home state requires for that receipt type)
• Product class: SaaS subscription, professional services, tangible goods, marketplace fee, implementation
• Amount, tax collected, net received, and processor fee
• Reconciliation key to Stripe / App Store / Paddle / bank deposit

Reconcile gross sales to 1099-K and bank deposits monthly. If your processor reports $620K gross but your ledger shows $598K net of fees, you need to book the $22K as processing expense, not as a revenue reduction, and apportion on the gross receipt figure the statute uses.

Tooling tip: if you use plain-text accounting, keep this as a structured journal with explicit state tags and product-class accounts, so your apportionment query is a single report, not a forensic exercise. See the patterns in /docs/ and the dashboards in /fava/ for slicing income by custom dimensions.

2. Map nexus state by state​

For each state where you have customers, answer:

• Do I have physical presence, employees, contractors, inventory, or owned delivery?
• Do I exceed that state's economic nexus threshold for income tax (often $100K sales, but varies)?
• Am I protected by P.L. 86-272 for that receipt type? For SaaS the answer is almost always no.

Mark every state as nexus / no-nexus for income tax purposes. This is separate from sales tax nexus, though the data overlaps.

3. Run the throwback vs. throwout math​

With the ledger and nexus map, calculate:

• Home-state sales factor without adjustment
• Factor with throwback (add nowhere sales to numerator)
• Factor with throwout (remove nowhere sales from denominator)

Do this for each state where you file as the home state, and for each state where you file as an out-of-state taxpayer that has a throwback rule — both can claim the same nowhere sale.

Save the three scenarios and the dollar impact at your marginal state rate. The delta is your exposure and your planning budget.

4. Track sourcing-rule changes as events, not footnotes​

When a state like Arkansas phases out throwback or Louisiana repeals throwout for intangibles, it alters your numerator or denominator the next tax year. Log these as dated entries in your tax calendar so your model uses the correct rule for the correct year. Annual review is not optional — the roster has shrunk by five states since 2019.

Five Practical Moves That Actually Reduce the Hit​

None of these are about hiding sales. They are about aligning operations with how states claim them.

1. Establish nexus intentionally where it helps​

If a throwback state is recapturing $2M of nowhere sales because you have no taxable presence in destination states, creating nexus in a low-tax or no-tax destination state does not by itself fix throwback — the destination must actually be able to tax you on that income. In many cases, voluntarily registering and filing where you have material sales converts nowhere sales into taxable-there sales that are no longer thrown back. Run the math before you register; sometimes paying a small tax in State B saves a large tax in State A.

2. Separate receipt streams by sourcing class​

Do not book SaaS, services, and tangible goods to one revenue account. States source them differently and apply throwback differently. Splitting them in your chart of accounts lets you apportion each stream under its correct rule and document why a SaaS stream was excluded from a tangible-property throwback. Auditors look for the workpaper.

3. Document market-based sourcing with defendable data​

For SaaS, keep the customer address that drove the sourcing, the contract language on where benefit is received, and any look-through for enterprise seats across states. If you allocate a $120K enterprise deal across 10 user locations, keep the allocation methodology. "We used billing address for all" is only defendable if that is what the statute allows.

4. Model Joyce vs. Finnigan if you have affiliates​

If you bill SaaS from an operating LLC owned by a holding company and file combined in a Finnigan state, the group's combined sales determine throwback. Filing separate or restructuring billing entities can change the outcome, but it has corporate, legal, and payroll consequences. Model both before you reorganize.

5. Time apportionment-factor investments​

Because single-sales-factor states reward in-state property and payroll lightly, moving engineers or servers does little to reduce apportionment there. But in three-factor states where sales are weighted heavily, even modest in-state sales shifts matter. If you are choosing where to hire or where to host infrastructure, the apportionment weight should be one input alongside talent and cost.

What to Keep in Your Books So an Audit Is Boring​

Auditors ask for the same package every time. Build it as you go:

• Monthly sales-by-state reconciliation: ledger vs. processor vs. bank vs. sales tax filings, with variance explanations
• Product-class mapping: which SKU maps to tangible vs. service vs. SaaS, and which sourcing rule and throwback rule you applied
• Nexus workpaper: threshold calculations and filing determinations by state, updated when receipts cross bright lines
• Apportionment model: the three-factor or single-factor calculation with and without throwback/throwout, saved by tax year
• Sourcing evidence: customer addresses, contract benefit language, allocation keys for multi-state enterprise deals

When this lives in version-controlled plain text, you get an audit trail for free — every change has a date, an author, and a reason. That is exactly what a state auditor wants to see when you claim a sale was market-sourced to State B and therefore not thrown back to State A.

Common Mistakes SaaS Founders Make​

"P.L. 86-272 protects my remote SaaS sales." It does not. It protects only tangible property where solicitation is your only activity. Services and subscriptions fall outside it entirely. Relying on it for SaaS is the single most common error.

"We have no inventory, so throwback doesn't apply." For pure SaaS in a tangible-only throwback state, that may be true this year — but add one hardware SKU, one conference merch run, or one on-prem appliance and the tangible throwback snaps on for that stream. Books that do not split streams cannot prove the exclusion.

"We use billing address for everything, everywhere." Some states require market-based on benefit received or look-through to user location, not billing address. Others still use cost-of-performance. One sourcing key does not satisfy all states.

"We file combined, so intercompany doesn't matter." It determines whether an affiliate's nowhere sales count under Joyce versus Finnigan. Intercompany eliminations and entity-by-entity nexus still matter.

"The state list never changes." It does, and recently toward repeal. If your model still assumes Alabama or West Virginia has throwback, you are over-provisioning; if it misses a new market-based adoption, you may be under-provisioning.

A Quick Checklist Before Your Next Filing​

• Do you have revenue tagged by destination state and by receipt class (tangible, SaaS, services)?
• Have you determined income-tax nexus separately from sales-tax nexus for each state?
• Have you identified which of your filing states impose throwback or throwout and for which receipt types?
• Have you calculated the home-state factor with and without the adjustment and converted the delta to dollars?
• Is your SaaS sourcing rule documented (market vs. cost-of-performance) with the underlying customer data retained?
• Did you check for combined-reporting and Joyce/Finnigan implications if you have more than one entity?
• Have you logged any 2024-2026 repeals or phase-outs that change next year's model?

If you can answer yes to all seven, you will not be surprised when a state reclaims nowhere income, and you will have the paper to argue the amount.

Keep Your Books Ready for Whatever States Claim Next​

Throwback and throwout rules exist because states disagree about who gets to tax a sale made from one state into another where the seller has no footprint. For SaaS and service businesses, the interaction of economic nexus, market-based sourcing, and these recapture rules means a dollar of revenue can be taxed at home, taxed at market, taxed twice, or recaptured as nowhere income depending on how you track and source it.

The antidote is not a clever theory at filing time. It is a ledger that already knows — by state, by product class, by sourcing rule — where every dollar was earned and why you sourced it there.

Beancount.io gives you that foundation with plain-text accounting that is transparent, version-controlled, and built for querying. No black boxes, no vendor lock-in, and AI-ready data when you need to model apportionment scenarios or hand an auditor a complete trail. Get started for free and make your next state filing the boring one.