brycewang-stanford Grant Budget Guide
A skill for preparing realistic, well-justified research grant budgets. Covers common budget categories across major funding agencies (NSF, NIH, ERC, EPSRC, NSERC), cost estimation methods, budget justification writing, common mistakes that trigger reviewer concerns, and strategies for handling budget cuts during negotiation.
Budget Categories
Standard Budget Structure
Most funding agencies use similar budget categories, though terminology and rules vary. Understanding the universal structure helps when applying to any funder.
Standard Grant Budget Categories:
1. PERSONNEL (typically 50-70% of total)
a. Principal Investigator (PI) salary + benefits
b. Co-PI salary + benefits
c. Postdoctoral researchers
d. Graduate research assistants (stipend + tuition)
e. Undergraduate assistants
f. Research technicians / lab managers
g. Administrative support (if justified)
2. EQUIPMENT (items > agency threshold, e.g., > $5,000)
a. Major equipment purchases
b. Fabrication costs
c. Computer hardware (if specialized)
Note: Standard office computers usually NOT allowed
3. TRAVEL
a. Domestic conference travel
b. International conference travel
c. Fieldwork travel
d. Collaboration visits
e. Per diem and accommodation
4. MATERIALS AND SUPPLIES
a. Lab consumables (reagents, chemicals, glassware)
b. Software licenses
c. Computing costs (cloud, HPC allocations)
d. Participant incentives / compensation
e. Printing, copying, publication charges
5. OTHER DIRECT COSTS
a. Subcontracts / subawards
b. Consultant fees
c. Human subjects costs (IRB fees, participant payments)
d. Publication fees (open access charges)
e. Specialized services (sequencing, imaging, surveys)
6. INDIRECT COSTS (F&A - Facilities and Administration)
a. Calculated as percentage of modified total direct costs
b. Rate negotiated between institution and government
c. Typical US university rate: 50-65% (on-campus)
d. Some agencies cap indirect costs
Cost Estimation
Personnel Cost Calculation
Estimating personnel costs:
PI salary:
Monthly salary = Annual salary / 12
Person-months requested = Calendar months of effort
Cost = Monthly salary x Person-months
Benefits = Cost x Fringe benefit rate
Example:
Annual salary: $120,000
Effort: 2 summer months (of 12 calendar months)
Salary cost: $120,000 / 12 x 2 = $20,000
Fringe benefits at 30%: $6,000
Total PI cost: $26,000 per year
Postdoc salary:
Check institution's postdoc salary scale
NIH NRSA stipend levels provide a common benchmark
Include annual raises (3-4% per year)
Include full benefits (health insurance, retirement)
Example (3-year project):
Year 1: $56,000 + 30% fringe = $72,800
Year 2: $57,680 + 30% fringe = $74,984
Year 3: $59,410 + 30% fringe = $77,233
Total: $225,017
Graduate student:
Stipend (12 months): varies by institution ($25,000-$40,000)
Tuition remission: often charged to the grant ($15,000-$50,000)
Health insurance: $2,000-$5,000/year
Total per student per year: $42,000-$95,000
Inflation and Multi-Year Projections
def project_costs(base_cost, years, inflation_rate=0.03):
"""
Project costs over multiple years with inflation.
Most agencies expect 3% annual inflation for personnel
and 2-5% for supplies depending on the field.
"""
yearly_costs = []
for year in range(years):
adjusted = base_cost * (1 + inflation_rate) ** year
yearly_costs.append(round(adjusted, 2))
return yearly_costs
# Example: $50,000 postdoc base over 4 years
# project_costs(50000, 4, 0.03)
# Returns: [50000.00, 51500.00, 53045.00, 54636.35]
Budget Justification Writing
Structure and Tone
The budget justification is a narrative document that explains WHY each cost is necessary. It should demonstrate that every dollar serves the scientific goals.
Budget Justification Template:
A. Personnel
Principal Investigator (Dr. Jane Smith): 2 calendar months of
effort per year ($20,000 salary + $6,000 fringe = $26,000/year).
Dr. Smith will oversee all aspects of the project, including
experimental design, data analysis, and manuscript preparation.
Her expertise in computational neuroscience (see Biographical
Sketch) is essential for the proposed machine learning analyses.
Postdoctoral Researcher (TBD): 12 calendar months of effort per
year ($56,000 salary + $16,800 fringe = $72,800 Year 1, with 3%
annual increases). The postdoc will conduct the primary
computational analyses, develop the novel algorithms described in
Aim 2, and train graduate students on the methodology.
B. Equipment
GPU Computing Server ($35,000): A dedicated server with 4 NVIDIA
A100 GPUs is required for training the deep learning models
described in Aim 2. Cloud computing was considered but would cost
approximately $45,000 over the project period based on estimated
compute hours, making a dedicated server more cost-effective.
C. Travel
Conference travel ($3,000/year): Funds for the PI or postdoc to
attend one domestic conference (estimated: $1,800 for
registration, airfare, hotel, and per diem) and one workshop
($1,200). Target venues include NeurIPS and the Computational
Neuroscience annual meeting, where we will present project
results and recruit collaborators.
D. Materials and Supplies
Cloud computing ($5,000/year): AWS credits for preliminary
analyses and data storage. Estimated 500 GPU-hours at $3/hour
plus 10 TB storage at $0.023/GB/month.
Software licenses ($2,000/year): MATLAB license ($500),
specialized neuroimaging analysis package ($1,500).
Agency-Specific Rules
Common Restrictions by Funder
NSF (National Science Foundation):
- PI salary: max 2 months from all NSF grants combined
- Equipment threshold: $5,000 per item
- No construction costs
- Indirect costs: full negotiated rate
- Mandatory cost-sharing: generally prohibited (with exceptions)
NIH (National Institutes of Health):
- Salary cap: currently ~$221,900 (check annually)
- Modular budgets: up to $250K direct costs/year (R01)
- Detailed budgets: required above $250K/year
- Equipment threshold: $5,000
- Indirect costs: full negotiated rate
ERC (European Research Council):
- 100% of eligible costs + 25% indirect flat rate
- No consortium required (single PI)
- Host institution must be in EU/associated country
- Personnel must be directly employed
EPSRC (UK Engineering and Physical Sciences):
- Full economic costing (fEC) model
- EPSRC pays 80% of fEC
- Directly incurred, directly allocated, and indirect costs
- Estates and indirect rates set by institution
Common Budget Mistakes
Red Flags That Concern Reviewers
Mistakes to avoid:
1. Padding the budget:
Problem: Inflated costs "just in case"
Fix: Estimate realistically, use published cost data
2. Insufficient personnel:
Problem: Ambitious aims but only 1 person for 5 years
Fix: Match staffing to the work plan realistically
3. Missing cost categories:
Problem: No travel budget but claims dissemination plan
Fix: Budget must align with proposed activities
4. Unjustified equipment:
Problem: Expensive equipment without explanation of need
Fix: Explain why existing institutional resources are inadequate
5. No inflation adjustment:
Problem: Same salary Year 1 through Year 5
Fix: Include 3% annual increases for personnel
6. Excessive foreign travel:
Problem: 4 international trips per year for one PI
Fix: 1-2 conferences per year is typical and defensible
7. Large consultant fees without justification:
Problem: $50,000 for unnamed consultant
Fix: Name the consultant, explain unique expertise,
provide hourly rate and estimated hours
8. Forgetting open access fees:
Problem: Many funders now mandate open access publication
Fix: Budget $2,000-$5,000 per paper for APC charges
Handling Budget Cuts
Negotiation Strategies
When your budget is reduced by the agency:
Strategy 1 - Reduce scope proportionally:
"With a 20% reduction, we will focus on Aims 1 and 2,
deferring Aim 3 to a subsequent proposal."
Strategy 2 - Reduce personnel:
"We will hire 1 postdoc instead of 2, extending the
timeline for Aim 2 by 6 months."
Strategy 3 - Find alternative resources:
"The PI's department has agreed to provide 1 month of
summer salary, reducing the budget request by $15,000."
Strategy 4 - Adjust equipment plans:
"We will use institutional HPC resources instead of
purchasing dedicated hardware, saving $35,000 but
adding $8,000 in computing time charges."
What NOT to do:
- Do not agree to do the same work with less money
- Do not cut travel to zero (undermines dissemination)
- Do not eliminate graduate student support (reviewers notice)
- Do not remove the budget justification details
A well-prepared budget demonstrates that the research team has thoughtfully planned the project, understands what resources are truly needed, and will be responsible stewards of public funds. Budget quality correlates with proposal quality in the minds of reviewers, and a sloppy budget can undermine an otherwise strong scientific narrative.