Negotiating an Indirect Cost Rate
Nonprofits routinely underrecover overhead on federal awards, not because the rules forbid it but because working out what your real rate is takes financial analysis nobody has time for. The 2024 rewrite of the Uniform Guidance raised the de minimis rate to fifteen percent, which changed the calculation for thousands of organizations. AI can do most of the preparatory work behind that decision in an afternoon, and it can also produce a confident number that falls apart the moment a federal negotiator asks how you got it.
The nonprofit starvation cycle has been described so often that the phrase has lost its sting, but the mechanism is straightforward and still operating. Organizations underreport what it costs to run themselves, funders come to expect artificially low overhead, and organizations then cannot afford the finance staff, technology, and management capacity that would let them report accurately. Every year the gap between reported and actual infrastructure cost widens a little.
Federal awards are one of the few places where this is genuinely fixable, because the rules are on your side. The Uniform Guidance explicitly requires federal agencies to accept a recipient's negotiated rate, and it provides a floor for organizations that have never negotiated one. The problem has never really been permission. It has been that determining your actual indirect cost rate requires a cost allocation analysis that a two-person finance team cannot fit into a normal year.
The stakes rose in 2024. OMB's revision to 2 CFR 200.414, effective for awards issued on or after October 1, 2024, raised the de minimis indirect cost rate from ten percent to fifteen percent of modified total direct costs. The same rewrite raised the portion of each subaward included in the MTDC base from twenty-five thousand to fifty thousand dollars. For an organization with several million in federal funding, those two changes together are worth real money, and many nonprofits have not revisited their approach since.
This article works through the actual decision: what the de minimis rate gives you and what it costs you, when negotiating your own rate is worth the effort, what the analysis behind a rate proposal involves, and where AI can compress weeks of preparation into days. It also covers the specific ways an AI-assisted rate analysis can go wrong in a way that surfaces later, during negotiation or during a single audit.
You Have Three Options, and Most Organizations Default Into the Worst One
Every nonprofit receiving federal funds is on one of three paths whether or not anyone has consciously chosen it. The first is charging nothing for indirect costs, which is depressingly common and which means your unrestricted revenue is subsidizing federally funded work. The second is electing the de minimis rate. The third is negotiating a rate with your cognizant agency.
Charging nothing is almost never correct and is often the result of a misunderstanding. Organizations do it because an early grant application had no indirect line, because a program officer once discouraged it, or because someone believed that showing low overhead would make the application more competitive. None of these is a rule. Forgoing recovery is a decision to fund federal program delivery out of donations, and the board should be making that decision deliberately if it is going to be made at all.
The de minimis rate is the practical answer for most small and mid-sized organizations. Any recipient without a current negotiated rate may elect up to fifteen percent of modified total direct costs, without submitting a proposal, without documentation of the underlying costs, and without a negotiation. It is available indefinitely, and the election is straightforward. The cost of this simplicity is that fifteen percent may be well below what you actually spend on infrastructure.
Negotiating your own rate is the option that unlocks the difference. Nonprofits with meaningful infrastructure frequently calculate true indirect rates well above the de minimis floor, and a negotiated rate applies across federal awards for the period it covers. The price is a formal proposal to your cognizant agency, supporting documentation, a negotiation process, and an ongoing obligation to submit renewals. That is a real commitment, and it is the right one considerably more often than the sector's behavior suggests.
Charging nothing
Almost never the right answer
Your unrestricted revenue subsidizes federally funded programs. If this is your position, it should be a documented board decision rather than an accident of how the first proposal was written.
De minimis, up to 15%
Simple, immediate, capped
No proposal, no supporting cost documentation, no negotiation, available to any recipient without a current negotiated rate. Applied to modified total direct costs. The tradeoff is that it may sit well below your real cost.
A negotiated rate
More work, more recovery
A formal proposal to your cognizant agency produces a rate agreement other federal agencies are required to accept. Ongoing renewal obligations apply, and the underlying allocation must be defensible.
The Base Matters as Much as the Rate
A rate is meaningless without the base it applies to, and this is where organizations most often miscalculate what a change is worth. Modified total direct costs is not the same as total direct costs. MTDC includes salaries and wages, fringe benefits, materials and supplies, services, travel, and, under the current rules, up to the first fifty thousand dollars of each subaward. It excludes equipment, capital expenditures, participant support costs, rental costs, tuition remission, scholarships and fellowships, and the portion of each subaward beyond the fifty thousand dollar threshold.
Those exclusions change the arithmetic substantially for certain kinds of organizations. A program whose budget is dominated by participant support payments, direct assistance, or large subawards has a much smaller MTDC base than its total budget suggests, which means a given percentage yields far less recovery than the headline number implies. Two organizations with identical federal revenue and identical rates can recover very different amounts.
The subaward threshold change deserves specific attention because it is easy to miss. Under the prior rules only the first twenty-five thousand dollars of each subaward counted toward MTDC. The 2024 revision doubled that to fifty thousand. For a pass-through entity making numerous subawards, this alone materially increases the base, and organizations that built their budget templates before October 2024 may still be calculating on the old figure. If your organization makes subawards, it is also worth reviewing our guide to compliance monitoring across sponsored projects, since the same relationships drive both questions.
Before you evaluate whether to pursue a negotiated rate, calculate what each percentage point is actually worth on your real MTDC base across your current award portfolio. That single number determines whether this project is worth doing. For some organizations the answer is tens of thousands of dollars a year, which justifies significant effort. For others the base is small enough that the de minimis election is clearly correct and the analysis stops there.
What sits outside the MTDC base
Excluded costs that shrink what a rate recovers
- Equipment and capital expenditures
- Participant support costs, which are significant in training and stipend programs
- Rental costs of space
- Tuition remission, scholarships, and fellowships
- The portion of each subaward above the fifty thousand dollar threshold
What the Underlying Analysis Actually Involves
A rate proposal is fundamentally a sorting exercise. Every dollar of organizational cost has to be classified as direct, indirect, or unallowable, and every classification has to be consistent, justified, and applied the same way across the organization. That last requirement is the one that trips people up: you cannot treat a cost as direct on one award and indirect in the pool. Consistency is a compliance requirement, not a preference.
The unallowable category is where organizations lose credibility fastest. Certain costs cannot be charged to federal awards at all, including fundraising, lobbying, bad debts, most entertainment, and fines and penalties. These must be excluded from the indirect pool entirely. They also have to be excluded from the base in the right way, so that unallowable activities still bear their share of overhead rather than shifting it onto federal programs. A proposal that leaves development department costs sitting in the indirect pool will not survive review, and finding it there suggests to a negotiator that the rest of the analysis deserves scrutiny too.
Shared costs require an allocation methodology, and each methodology needs a rationale. Occupancy is typically allocated on square footage, information technology on headcount or device count, and administrative salaries on effort. What matters is not which basis you pick but that the basis reasonably reflects how the cost is actually consumed and that you can explain it. This connects directly to your time and effort records, which are the evidentiary foundation for any effort-based allocation in the proposal.
Then there is the reconciliation requirement. Your proposal has to tie back to audited financial statements. Every figure needs a traceable path from the audited totals through your allocation to the rate you are proposing. Negotiators check this, and a proposal that does not reconcile cleanly stalls immediately regardless of how sound the underlying logic is.
Finally, you choose a rate type. Provisional rates are temporary and later finalized against actuals. Predetermined rates are fixed for a period and not subject to adjustment. Fixed rates with carry-forward reconcile the difference between estimated and actual into a future period. Each carries different cash flow and risk characteristics, and the right choice depends on how stable your cost structure is and how much variance you can absorb.
What a proposal package contains
Broadly consistent across agencies
- Audited financial statements for the base year
- A schedule reconciling the proposal to those statements
- Personnel cost detail supporting effort-based allocations
- A written description of your allocation methodology
- A certification of the proposal by an authorized official
What gets a proposal sent back
Recurring reasons for delay
- Fundraising or lobbying costs left in the indirect pool
- Figures that do not reconcile to the audited statements
- Costs treated as direct on some awards and indirect in the pool
- Allocation bases asserted without a stated rationale
- Effort-based allocations with no effort records behind them
Where AI Compresses the Work
The reason most nonprofits never pursue a negotiated rate is not that the concept is hard. It is that the preparatory analysis is a large, unglamorous data project that competes with closing the books and getting reports out. AI changes the economics of that project substantially, particularly in the exploratory phase before you commit.
The most valuable early use is the feasibility estimate. Working from your general ledger and audited statements, a model can produce a first-pass classification of expense accounts into direct, indirect, and unallowable, calculate an approximate rate, estimate your MTDC base across current awards, and tell you roughly what the difference between that rate and fifteen percent is worth annually. That estimate is not a proposal and should never be treated as one, but it answers the only question that matters at the outset, which is whether this is worth pursuing at all.
The second is scenario modeling. Once you have a working classification, testing alternatives is fast: what happens if occupancy is allocated on square footage rather than headcount, what the rate looks like using a different base year, how the number shifts if a large one-time expense is excluded, how sensitive the result is to a single senior staff member's effort split. Doing this by hand in spreadsheets is slow enough that most organizations test one scenario. Testing eight changes the quality of the decision.
The third is drafting the methodology narrative. A proposal requires written explanations of why each allocation basis was chosen, and these are tedious to write and highly structured. Drafting them from your actual decisions, then editing for accuracy, saves meaningful time. The same applies to the internal memo explaining the recommendation to your board or finance committee, where the audience needs the reasoning translated out of accounting language.
The fourth is review before submission. Before anything goes to a cognizant agency, a structured check against the known failure patterns is worth running: unallowable costs in the pool, inconsistent treatment across awards, allocation bases without rationale, reconciliation gaps, effort assertions without records. This is the same detective posture that works well in audit preparation, applied earlier in the process.
One more use worth naming: monitoring after you have a rate. Rate agreements expire, renewals are due within a set window after your fiscal year end, and organizations lose recovery simply by missing the deadline. Tracking those dates alongside your other award obligations is exactly the kind of thing a grant compliance calendar should be handling.
A feasibility check you can run in a week
Enough to decide whether to commit to the full project
- Classify every expense account as direct, indirect, or unallowable, and have finance correct it
- Calculate an approximate rate from the corrected classification
- Compute your real MTDC base across all current federal awards
- Multiply the difference against fifteen percent to get annual dollars at stake
- Compare that figure against the cost of preparing and maintaining a proposal
- Take a documented recommendation to the finance committee either way
How an AI-Assisted Rate Analysis Goes Wrong
The failure mode here is specific and worth understanding, because it does not look like failure. An AI-assisted analysis produces clean schedules, consistent formatting, and a confident rate. What it cannot do is know things about your organization that are not in the data you gave it, and cost allocation is full of exactly that kind of knowledge.
Account name classification is the most common trap. A model sorting your chart of accounts works from labels, and nonprofit account labels are frequently wrong or ambiguous. An account called "Program Supplies" that in practice holds office supplies for the whole organization will be classified as direct, and nobody will notice unless a person who knows the ledger reviews the classification line by line. Every classification decision needs human confirmation, and this is the single most important review step in the process.
Split positions are the second. The development director who spends a quarter of their time on grant reporting, the program director who supervises across three awards, the finance manager who does both bookkeeping and program budgeting: these splits are not visible in the general ledger and they materially change the analysis. Only people who know the roles can supply them.
Third is the invented justification. Asked to explain why an allocation basis was chosen, a model will produce a fluent rationale whether or not that was your actual reasoning. If your organization allocates occupancy on headcount because that is what the previous finance director set up, the honest answer is that headcount reasonably approximates space consumption, not a constructed narrative about a space study you never performed. Negotiators ask follow-up questions, and a fabricated methodology does not survive them.
Fourth is stale regulatory detail. Models trained on older material will confidently cite a ten percent de minimis rate and a twenty-five thousand dollar subaward threshold, both of which were correct before October 2024 and are wrong now. Verify every regulatory figure against the current text of the rule rather than accepting it from a model, and treat any citation to a specific section as something to check rather than to trust.
Finally, a practical caution about data handling. A rate analysis involves your complete general ledger and detailed personnel cost information. Before any of that goes into a tool, confirm your organization's position on where financial data may be processed and retained. Our overview of privacy risk assessment for nonprofit AI projects covers how to make that determination.
Making the Call
The de minimis election is the right answer for a large share of nonprofits and there is no shame in it. If your calculated rate lands anywhere near fifteen percent, take the de minimis and spend the effort elsewhere. The negotiation, the annual renewals, and the ongoing obligation to maintain a defensible allocation are real costs, and they are not worth incurring for a marginal gain.
Pursue a negotiated rate when the gap is substantial and durable. Substantial means the annual dollar difference clearly exceeds what preparation and maintenance will cost you. Durable means your cost structure is stable enough that the rate will still be roughly right in two years, and that your federal funding is likely to continue at a similar level. Organizations in the middle of major program changes or facing uncertain federal funding may reasonably wait.
Whichever way you go, the analysis is worth doing. Knowing your real indirect cost rate is useful well beyond federal awards. It tells you what to ask foundations for, what a new program actually costs to run, and whether your fee-for-service pricing covers its overhead. Many organizations discover during this exercise that their true rate is far higher than what they have been requesting from every funder, federal or not, which is a finding worth having regardless of what you do about the federal piece.
Bring the result to your board. A finance committee that understands the organization's real infrastructure cost is better equipped to push back on funders who expect overhead to be invisible, and better able to explain to donors why administrative cost is not waste. That conversation is one of the more useful things a well-prepared rate analysis produces, and it happens whether or not you ever submit a proposal.
Conclusion
Indirect cost recovery is one of the few places in nonprofit finance where money is being left on the table for procedural rather than substantive reasons. The rules permit recovery, the de minimis floor has risen, and the main barrier has always been that the analysis required to know your own number was too expensive to perform.
That barrier has genuinely come down. A feasibility analysis that used to be a consulting engagement is now something a finance team can run in a week, which means the decision can be made on evidence rather than on inherited habit. What has not changed is that the number has to be right and the reasoning has to be yours. AI can classify, calculate, model, and draft. It cannot know that the account labeled program supplies is really office supplies, and that single fact is the difference between a proposal that gets approved and one that gets sent back.
Run the feasibility check. If the answer is fifteen percent, elect it and move on with confidence. If the answer is meaningfully higher, you now have a documented case for capturing the difference, and the analysis you built to get there will keep paying off in every funder conversation you have afterward.
Find Out What Your Real Overhead Rate Is
We help nonprofit finance teams use AI on the analysis that has always been too expensive to do, without cutting corners on the parts that have to be defensible.