Traffic Impact Study vs. Feasibility Study: What Each One Answers and How SBA and USDA Lenders Use Both
September 25, 2026 by Michal Mohelsky, principal of MMCG Invest, LLC
A traffic impact study tells a highway agency what a project will do to the road. A feasibility study tells a lender what the road will do for the project. Neither SBA nor USDA asks for the first one by name, and both end up underwriting its conclusions anyway.
The same vehicle counted twice
This summer we read a comment-and-response letter between a traffic engineer and a state highway district about a proposed fuel and convenience store at the signalized junction of two state routes in the Midwest. Nobody would mistake it for literature. The district reviewer had seven comments. The labels for the two state routes were reversed on every figure. The pass-by trips for northbound traffic looked wrong on a dozen exhibits, and the engineer wrote back that the volumes were correct and that the reviewer had probably missed a through movement at the third intersection. The capacity results had to be recast into the state's own table layout, with cycle lengths, volume-to-capacity ratios and 95th percentile queues. The study had already concluded that the site needs a new southbound left-turn lane at its west driveway, so the reviewer wanted storage lengths calculated for that lane, for a proposed westbound left at the north driveway, and for every existing left turn at the signal, where the bays run 190 to 230 feet today. Left-turn phasing at the signal had to be tested against the state traffic engineering manual, with crash counts by approach. And the reviewer had not opened the capacity analysis files at all, because the volumes were in dispute. (1)
The engineer's closing paragraph asks for a prompt review "so construction may begin." A lender should read that sentence twice. The access permit, the building start, the draw schedule and the first interest payment on this project all sit behind an argument about pass-by trips.
Here is why the episode deserves more than a shrug. The pass-by trips the engineer and the reviewer are arguing about are, vehicle for vehicle, the customers in the borrower's revenue projection. The traffic study subtracts them, because a car that was already on the road adds nothing to the road. The feasibility study counts them, because a car that was already on the road is the cheapest customer a gas station will ever get. One document treats the passing stream as a burden to be discounted. The other treats it as the business.
Borrowers confuse the two studies constantly, and some lenders do too. We get calls asking whether the traffic study "covers" the feasibility requirement, and calls asking whether our feasibility study can be handed to the DOT. The answer to both is no, and the reasons are worth laying out carefully, because the two documents collide in the credit file whether or not anyone planned for it.
The argument of this article, stated up front:
SBA does not require a traffic impact study anywhere in SOP 50 10 8.1. USDA does not require one anywhere in 7 CFR Part 5001. Neither agency uses the phrase. Yet the conclusions of a traffic study enter an SBA file through four doors and a USDA file through three, and they change the three numbers a guaranteed lender cares about most: total project cost, time to certificate of occupancy, and debt service coverage at stabilization.
The traffic study and the feasibility study also contain one assumption in common, written in two dialects. The engineer calls it the pass-by percentage. The analyst calls it the capture rate. They are arithmetically tied to each other, and a credit file in which they disagree has an error in it somewhere. We show the arithmetic below.
Two documents, two clients, two questions
A traffic impact study is an engineering report written for a road agency. The developer pays for it, but its real reader is the state DOT district office or the municipal traffic engineer who decides whether a driveway or encroachment permit gets issued and on what conditions. The study estimates how many vehicle trips the project will generate, using rates published in the Institute of Transportation Engineers Trip Generation Manual. (2) It distributes those trips across the surrounding network and tests the affected intersections for delay, level of service and queuing under Highway Capacity Manual methods. (3) Its conclusions are physical. Build this lane. Lengthen that turn bay. Restrict this driveway to right turns. Retime that signal.
A feasibility study is a market and financial report written for a credit decision. The borrower usually pays for it too, but its reader is the loan committee and, behind the committee, the agency guaranteeing the loan. It estimates demand, revenue, operating cost and the cash available to service debt. Its conclusions are financial: whether the project, at this cost and in this market, can carry this loan. Under SBA's regulations it is a document the agency "may require." (4) Under USDA's it is a defined term with five mandatory components and an independent author. (5)
The two studies share exactly one input, the traffic count on the frontage road, and they part ways immediately after. The traffic engineer is professionally wary of trips. Every vehicle is load on an intersection, and the job is to prove the network can absorb it or to specify what must be built so that it can. The feasibility analyst is looking for demand. Every vehicle is a prospective transaction, and the job is to estimate, defensibly, how many of them turn in. Neither posture is wrong. They are answers to different questions, asked by different institutions, with different consequences for getting it wrong. An engineer who understates trips gets a failed intersection. An analyst who overstates customers gets a defaulted loan.
The timing differs as well, and this is where deals get hurt. The feasibility study is normally ordered early, when the lender is forming a view. The traffic study is scoped during entitlement and permitting, often after the term sheet, sometimes after the budget has been fixed. When the DOT's conditions arrive late, they arrive as changes to a project the lender thought it had already underwritten.
What SBA says about traffic studies: nothing, and then a great deal
SOP 50 10 8.1 governs 7(a) and 504 loans numbered on or after October 1, 2026. (6)(7) It does not mention a traffic impact study, a traffic study or a traffic impact analysis in either program. Eligibility turns on operating status, size, business type, ownership and credit elsewhere. The nearest thing to a location rule is occupancy: 51 percent of an existing building, 60 percent of new construction. Nothing in the eligibility chapter cares whether customers can turn left into the site.
The nearest codified hook is one permissive sentence in the regulations: "SBA may require professional appraisals of the applicant's and principals' assets, a survey, or a feasibility study." (4) We have written at length elsewhere about what that sentence does and does not do (see "The SBA Feasibility Study Requirement That Does Not Exist"). For present purposes the point is narrow. A survey and a feasibility study are the two instruments in the SBA framework that sit closest to a traffic analysis, and both are discretionary.
So the SOP is silent on the study. It is not silent on the things the study produces. Those come in through four doors.
The first door is eligible project cost. On the 7(a) side, proceeds may be used to "improve a site (e.g., grading, streets, parking lots, landscaping), including up to 5 percent for community improvements such as curbs and sidewalks." (8) On the 504 side, land improvements integral to the project are eligible costs, and the SOP's examples read like a traffic engineer's punch list: grading, new streets including curbs and gutters, parking lots, utilities. Professional fees tied to zoning, permits and platting are eligible as well. (6)(9)
On-site access work fits these provisions without strain. Driveways, internal circulation, curb cuts and parking are ordinary site improvements. The harder case is off-site work, which is exactly what a DOT permit tends to order. A left-turn lane on a state route is built in public right-of-way, on land the borrower does not own, and is dedicated to the state when finished. It never becomes collateral. That does not make the cost ineligible. In 504, the test is whether the improvement is integral to the project and whether it is being paid for through a special tax assessment or user fee, in which case it is excluded. A turn lane the borrower must build and pay for directly as a condition of the permit is about as integral as a cost gets, since the project cannot legally open without it. The same lane financed through an assessment district falls outside eligible project cost. In 7(a), the community improvement slice carries the explicit 5 percent ceiling. (8)(9) The SOP supplies the test and the lender or CDC applies it to the deal. What a lender should take from this is simple: the lane is financeable in most structures, which means it will be in the loan, which means it will be in the debt service.
The second door is the contingency and the schedule. Procedural Notice 5000-872764 raised the 504 construction contingency ceiling from 10 to 15 percent of construction cost in September 2025, and the change carries into 8.1. (10) The underlying regulation still reads 10 percent, an inconsistency SBA has not yet cleaned up, and a separate provision governs what happens to unused contingency at closing. (9)(11) There is no codified 7(a) contingency percentage at all. Either way, a contingency is sized against the budget as the lender understood it. A $250,000 turn lane that surfaces in a DOT review letter four months after commitment can consume most of a contingency before a footing has been poured.
The third door is the feasibility study itself, when the lender orders one. The June 2025 rewrite of the SOP made coverage tests prescriptive, with a floor of 1.15 to 1 on standard 7(a) loans, and the practical effect has been to raise the burden of proof on any file that rests on projections: startups, ground-up construction, special-purpose property. (6) A feasibility study on a gas station, a car wash or a drive-through restaurant that does not address whether vehicles can physically reach the pumps is defending its revenue line with one hand tied. Access conditions belong in the market analysis because they govern capture, in the site analysis because they govern operation, and in the cost build-up because they govern the budget.
The fourth door is the appraisal and the environmental review. For construction loans the SOP requires value at completion, and an appraiser working through highest and best use will notice a right-in, right-out restriction or a missing median break and price it. A corner site with full movement and a mid-block site that can only be entered from one direction are not comparable, and a competent appraisal will say so. The Phase I environmental site assessment, mandatory for gas stations regardless of loan size, reviews adjoining properties and historical uses along the frontage, which is how roadway corridor conditions occasionally surface there too. (6)
Four doors, no mention of the study. An SBA lender can close a construction loan without ever seeing the traffic study, and many do. The lender cannot close one without underwriting its consequences, because they are already in the budget, the schedule, the projections and the appraisal.
What USDA says: a codified study with roads written into it
USDA is a different animal, and the difference matters for anyone who works in both programs. The OneRD guaranteed loan regulation at 7 CFR Part 5001, which covers Business and Industry, Community Facilities, REAP and Water and Waste, defines a feasibility study as "a report including an opinion or finding conducted by an independent qualified consultant(s) evaluating the economic, market, technical, financial, and management feasibility of the proposed project." (5) That one sentence does two things SBA's rules never do. It fixes the content, five components, and it fixes the author, who must be independent and qualified.
Appendix A to Subpart D fills in the components. (12) Technical feasibility is described as the analysis of "the delivery of goods or services, including transportation, business location," and the factors the appendix lists beneath it include "Roads, rail, airport infrastructure" and "Need for local transportation." Market feasibility calls for sales or service estimates covering current and prospective buyers or users, which for a roadside business is a statement about who can see the site and reach it. The words "traffic study" do not appear. The subject matter of one is written directly into the required content of the feasibility study.
The study is also mandatory in defined cases, not discretionary. A B&I guaranteed loan above $1 million to a new business requires one from an independent qualified consultant. (13) Community Facilities loans above $1 million to a new entity, or to an existing one starting a new activity, require one as well. (14) The regulation is otherwise sparing with hard numbers; the recency rule lenders quote most often, 90 days, applies to the borrower's financial statements and not to the study's market data. (15)
Traffic has a second way into a USDA file, and the citation for it changed this year. Rural Development's environmental reviews ran under 7 CFR Part 1970 for a decade. That part is gone. USDA consolidated its agency NEPA rules into a single departmental regulation at 7 CFR Part 1b, the final rule was published on April 3, 2026, and Rural Development confirmed in May that all applications must now follow Part 1b. (16)(17) Anyone still citing Part 1970 in a study or an application narrative is citing a removed regulation.
Part 1b is where traffic is named outright. The categorical exclusion for small-scale site development applies to projects that would disturb "not more than 10 acres of real property and would not cause a substantial increase in traffic." (18) A project that would cause one loses the exclusion and moves up to an environmental assessment, with the time and cost that implies. For a rural travel center or a distribution facility on a two-lane county road, "substantial increase in traffic" is not a hypothetical. It is the project description.
The third way in applies to Community Facilities and water projects, through the Preliminary Engineering Report. The PER outline in RUS Bulletin 1780-2 has no line item for roads, traffic or ingress. Access shows up sideways, in the requirement to identify sites, easements and access agreements, and in a list of potential construction problems that includes "limited access." (19) REAP projects, for their part, run on a tiered technical report focused on the energy resource and interconnection, with a feasibility study required only for renewable energy systems when the lender or the agency considers it necessary. (20)
Set the two agencies side by side and the contrast is sharper than most practitioners assume. SBA has no content standard for a feasibility study, no preparer standard, and no requirement that one be produced. USDA has all three. Neither agency names a traffic study. But a USDA feasibility study that ignores roads and access has omitted an enumerated factor, while an SBA feasibility study that ignores them has merely done a poor job. One is a compliance defect and the other is a credit defect. A lender would prefer to have neither.
When a traffic impact study is triggered
Nobody orders a traffic study for the pleasure of it. It is triggered, usually by a number, and the number depends on who owns the road.
The rule of thumb across the industry is 100 new vehicle trips in the peak hour. That figure is real, but it is mostly a municipal figure. Portland's code, for example, calls for a study at more than 100 new peak-direction trips in the site's peak hour, or more than 250 new trips a day. (21) State DOTs, which control access to state routes, often set their own lines elsewhere.
Georgia is explicit: the Department "will require a traffic impact study for any site estimated to generate more than 2,000 gross daily trips," and reserves the right to require one below that. (22) North Carolina's threshold is 3,000 daily trips, and NCDOT says plainly that many municipalities are stricter and that the trip count is only one factor in the decision. (23) Ohio's State Highway Access Management Manual scales the analysis to the volume, with a lighter access study for developments under 100 peak-hour trip ends and a full impact study above that, and it treats a change of use that raises driveway volume by 20 percent as a fresh trigger. (24) Texas describes "minor engineering studies" below 100 trips and then declines to commit, saying the need for a full analysis is determined by TxDOT where it is the permitting authority. (25) Florida's Site Impact Handbook emphasizes capacity and safety impacts over a fixed count, which in practice leaves the numeric trigger to counties and cities. (26)
California needs its own paragraph because it is widely misunderstood. Senate Bill 743 and CEQA Guidelines section 15064.3 replaced automobile delay with vehicle miles traveled as the measure of a project's transportation impact under CEQA. (27) Many developers heard that as the end of the traffic study. It was not. The statute changed the environmental document. It did not change the encroachment permit, and ITE's own guide to the law notes that delay and level of service remain in use for intersection design, signal warrants and operational analysis. (28) A California fuel site on a state highway now needs both: a VMT analysis for CEQA and an operational study for Caltrans.
Two practical observations follow from all of this. First, the stricter of the state and local thresholds governs, and a site with state route frontage gets two reviewers. Second, the asset classes that fill SBA and USDA pipelines sit on opposite sides of the line. A 100-room hotel generates about 60 trips in the evening peak hour and a self-storage facility generates almost none. A twelve-pump convenience store generates several hundred. Fuel, quick-service restaurants, coffee drive-throughs and car washes will trip nearly any threshold in the country. Storage and small hotels frequently will not.
What the study contains, what it costs, and how long it takes
The scope of a traffic impact study is remarkably consistent from state to state. A scoping letter fixes the study area, which usually means the site driveways and the nearest significant intersections. The engineer collects turning movement counts, normally no more than a year or two old. Trips are generated from ITE rates, reduced for pass-by and internal capture under the methods in the ITE Trip Generation Handbook, then distributed and assigned to the network. (29) The analysis is run for existing conditions, a future no-build case and a future build case, in the morning and evening peaks, for the opening year and a design year. Each intersection gets delay, level of service, volume-to-capacity and 95th percentile queue results. (3) Then come the warrant analyses: whether left-turn and right-turn lanes are justified under the AASHTO Green Book and the NCHRP Report 457 methodology that most state manuals adopt, and whether a signal is justified under the nine warrants in Chapter 4C of the Manual on Uniform Traffic Control Devices. (30)(31)(32)
The trip generation table is the heart of the document, and it is worth seeing the rates for the property types lenders actually finance. The figures below come from the 11th Edition of the ITE manual, which is the edition most state manuals and studies in circulation still reference. (2) A convenience store with fuel generates roughly 18 to 27 trips per fueling position in the evening peak hour, depending on the size of the store. A fast-food restaurant with a drive-through generates about 33 per thousand square feet. A coffee shop with a drive-through runs higher still, around 39 to 44. An automated car wash generates about 78 per tunnel. A hotel generates 0.60 per room, and a self-storage facility 0.15 per thousand square feet, which is to say almost nothing. These are averages with wide ranges around them, and several of them rest on small samples that ITE itself flags for caution.
On cost, traffic engineers quote by the intersection. One firm's published rule of thumb is $4,000 per intersection, ranging from $3,000 to $5,000 when three or more are studied. (33) Another puts it at $3,500, with a range of $2,500 to $5,000. (34) A third describes full studies as typically running $10,000 to $30,000. (35) These are not contradictory. A rural site with two driveways and one signal is a four-figure study. A corridor site with six intersections, two agencies and a design-year horizon is a five-figure one. For comparison, and because borrowers conflate the two fees as well as the two documents, a lender-grade feasibility study is a separate engagement with a separate fee and a different author.
Time is the more expensive variable. Scoping and agency concurrence on methodology commonly take a few weeks. Production takes three to eight. Portland's transportation bureau tells applicants to plan on at least seven weeks for the initial review alone. (36) Then the comment letter arrives, as it did on the Midwestern site, and the clock restarts. Figures get corrected, tables get reformatted, a disputed volume gets defended or conceded, and the package goes back in. Two rounds is ordinary. Three is not unusual when the engineer and the reviewer disagree on a technical point, and in our example they disagreed on the central one. (1) Four to six months from scoping to an issued access permit is an unremarkable outcome.
For a lender, the way to think about those months is in interest cost. A construction loan with $2 million drawn at 9.5 percent costs about $15,800 a month to carry. A project that has closed on land and is waiting on an access permit is paying that carry, or its equivalent in equity cost, for every round of review.
What the study orders you to build
A traffic study's recommendations are not advice. Once the agency accepts them they become permit conditions, and the permit is what allows the driveway to exist. The common conditions are a left-turn lane with a specified storage length, a right-turn deceleration lane, a two-way left-turn lane along the frontage, a new signal, or modifications to an existing one.
Published cost data for these items is scattered, regional and mostly a few years old, so what follows should be read as ranges with their sources attached and not as estimates for any particular site.
A review of TxDOT lettings by the Texas A&M Transportation Institute found left-turn lane projects running from $120,000 to $400,000 in construction cost, with an average near $250,000. (37) Arizona DOT's safety program pricing examples include a lighted left-turn lane on a state route that bid between $533,000 and $570,000 and totaled $708,889 once design and indirect costs were included. The same data set shows three right-turn lanes on another state route totaling $543,981, or roughly $180,000 a lane all in. (38) Storage length is the cost driver people overlook. Every additional 25 feet of bay is more earthwork, more pavement and more drainage, and sometimes more right-of-way, which is why the DOT's queue calculation in the comment letter is a budget item dressed as an engineering detail.
Signals are dearer. Washington State DOT says a basic signal system costs around $250,000 for the installation alone and closer to $500,000 or more as a complete project with engineering and traffic control, and it puts annual operating cost at about $8,000. (39) North Carolina's DOT quotes $50,000 to over $100,000, a figure that reads like hardware only. (40) Arizona's examples start at $405,937 for a three-leg signal and climb steeply with complexity. (38) Modifying an existing signal is cheaper but poorly documented. Retiming alone runs $2,000 to $3,100 an intersection. (41) Adding a protected left-turn phase can mean new signal heads, new detection and sometimes new mast arms, at which point it is a partial rebuild. Frontage widening is priced by the mile: Florida's conceptual cost models put the increment for a center turn lane on a rural road at roughly $475,000 a mile. (42)
Readers of our earlier article on site risk (SBA Approves Borrowers. Sites Kill Loans.) will remember a lower figure for a single deceleration lane, from about $30,000 into the low $100,000s. Both numbers are honest. The lower one describes a short lane on a simple frontage built by the site contractor as part of the sitework. The higher ones describe work let to DOT standards with drainage, lighting, design fees and traffic control. Which one a project gets depends on the road, and the borrower rarely gets to choose.
Now put one of these into a loan. Take a fuel and convenience project with stabilized EBITDA of $620,000 and annual debt service of $480,000, a coverage ratio of 1.29 times. The access permit comes back requiring a $250,000 left-turn lane. At 80 percent loan-to-cost, $200,000 of that is new debt, which at current SBA pricing over 25 years adds about $21,000 a year in payments. Coverage falls to 1.24 times. That is survivable. It is also a third of the distance to the 1.15 floor, consumed by a single line item that did not exist when the feasibility study was written. Add four months of delay and the carry that goes with it, and a comfortable file has become a thin one without a single customer behaving differently than projected.
The customers are the subject of the next two sections.
Pass-by and capture rate are the same number seen from two sides
ITE sorts the trips at a site's driveway into three kinds. A primary trip is made for the specific purpose of visiting the site. A pass-by trip is made by a driver already on the adjacent road, on the way somewhere else, who stops without changing route. A diverted trip comes from a driver on a nearby road who detours to reach the site. (29) Pass-by trips add no vehicles to the surrounding network, so the engineer deducts them before testing the intersections. The deduction is large for roadside businesses. ITE's average evening peak pass-by for a convenience store with fuel is 56 percent, and Wisconsin DOT's compilation of the underlying studies shows a range of 46 to 72 percent. (43) For fast food with a drive-through the published averages run from 49 to 55 percent, with individual sites anywhere from 25 to 71. (43)(44) ITE publishes no pass-by rate for coffee drive-throughs, so engineers borrow the fast-food figure. Self-storage is not eligible for a pass-by reduction at all. Nobody rents a storage unit on impulse.
A feasibility study approaches the same vehicles from the other direction. It starts with the traffic count on the frontage and applies a capture rate, the share of passing vehicles expected to turn in, adjusted for visibility, access, corner position and the direction of the commute. Capture rates are the softest number in roadside retail analysis. There is no published standard. The figures in circulation are practitioner heuristics and vendor model outputs, and an analyst under pressure to make a deal work can move one by a point or two without anyone noticing.
Except that the traffic study is sitting in the same file, and it contains a check on exactly that number. If pass-by visits are the customers who come from the passing stream, then:
capture rate × daily traffic on the road = pass-by share × daily visits to the site
Rearranged, the capture rate implied by the traffic study equals the pass-by share times the site's daily visits, divided by the road's daily traffic.
Work it through. NACS reports that the average U.S. convenience store handled 1,484 transactions a day in 2025. (45) Treat that, roughly, as 1,500 vehicle visits for a site on a road carrying 20,000 vehicles a day. If 56 percent of those visits are pass-by, 840 come from the passing stream and the other 660 are primary and diverted trips drawn from the surrounding trade area. The implied capture rate is 840 divided by 20,000, or 4.2 percent.
Now suppose the feasibility study for that site assumes a 7 percent capture rate. That is 1,400 visits a day from the stream alone. If the study also credits the site with the 660 trade-area visits, total volume reaches 2,060 a day, about 39 percent above the national average store, on a road that barely clears the volume fuel operators consider a minimum. If instead total visits are held at 1,500, a 7 percent capture rate implies that 93 percent of customers are pass-by, far outside the range ITE has ever observed. Either the capture rate is too high, the traffic study's pass-by share is too low, or the site is genuinely exceptional and the study should explain why. What the file cannot sustain is both documents being right.
The same identity exposes a quieter inconsistency. A developer sometimes argues to the DOT that the site's trips are overwhelmingly pass-by, to minimize required mitigation, while the pro forma argues to the lender that the site will pull new destination demand from a wide trade area. Those are opposite claims about the same cars. The peak-hour math makes it concrete: twelve fueling positions at ITE's middle rate of 22.76 produce 273 evening peak trip ends, of which 153 are pass-by and 120 are new. The 120 drive the mitigation. The 153 drive the revenue. A borrower cannot shrink the first number for one audience and the second for another.
When the permit restricts the driveway
Cost and delay are the visible consequences of a traffic study. The expensive one is quieter. When an agency concludes that left turns into or out of a driveway are unsafe, it conditions the permit on right-in, right-out operation, or it closes a median opening, or it replaces a two-way left-turn lane with a raised median. Each of those is a traffic engineering decision. Each is also a statement about how much of the passing stream can still reach the site.
The research on this question is better than most owners expect and more reassuring than most owners believe. NCHRP Report 420 remains the foundational survey of access management techniques and their effects. (46) FHWA's 2024 summary of the evidence concludes that after construction, access changes are unlikely to harm a business, that the effects which do occur are concentrated in the construction period, and that retail sales typically return to prior levels or rise. (47) An Iowa State study of access-managed corridors found that only 5 percent of businesses reported lower sales afterward, against 33 percent reporting higher sales and 53 percent no change. (48) Utah's DOT used state taxable sales data, not opinion surveys, and found that owners on median corridors perceived far more damage than the receipts showed. (49) The safety case is not in dispute at all.
If the article stopped there it would be misleading, because the same literature carves out an exception, and the exception is our clients. State DOT surveys summarizing these studies report little or no effect on businesses generally, with the exception of impulse-oriented ones. The Texas Transportation Institute's study of a median project in College Station found most business types reporting more customers afterward, while the gas stations in the sample reported gross sales down 17.6 percent during construction and 40.4 percent after it. (50) That sample was five stations reporting their own impressions, and nobody should put 40 percent in a model on the strength of it. But the direction is consistent with everything site-selection practitioners say about fuel and quick-service sites: the customer is making a decision in a few seconds at 45 miles an hour, and a customer who has to pass the site, make a U-turn and come back mostly will not. The working range practitioners use for the difference between full-movement and right-in, right-out access is 15 to 30 percent of capture. It is a heuristic, not a published standard, and we treat it as the boundaries of a stress test and nothing more.
Run that stress test on the example from the previous section. The site draws 840 visits a day from the passing stream. A restriction that costs it 15 percent of that capture removes 126 visits a day, 8.4 percent of total volume. Apply that to $1.9 million of gross profit, allow a quarter of the loss to be recovered through variable cost savings, and EBITDA falls by about $120,000, from $620,000 to $500,000. Coverage drops from 1.29 to 1.04 times. At a 30 percent loss of capture, EBITDA falls to about $381,000 and coverage to 0.79. A convenience store's cost structure is mostly fixed, so a single-digit loss of customers becomes a double-digit loss of cash flow. The turn lane in the previous section cost five basis points of coverage. The driveway restriction costs twenty-five to fifty.
This is the collision the title of this article refers to. The traffic study records the restriction as a safety improvement with no effect on the network, which is correct, because the pass-by trips it eliminates were never new trips to begin with. The feasibility study has to record the same restriction as a revenue event, because those trips were the business. A file that contains the first document and a feasibility study written before it, or written without reference to it, has an unpriced risk sitting in plain view.
What a lender should ask for
None of this argues for making a traffic study a condition of every guaranteed loan. For a self-storage facility or a 70-room hotel it would usually be a waste of the borrower's money. It argues for something narrower: on traffic-dependent property with state route frontage, the lender should know the status of the access permit before it relies on the projections, and the feasibility study should be written with the traffic study in hand or, where it does not exist yet, with its likely conclusions explicitly assumed and stressed.
In our own engagements that comes down to a handful of habits. We ask for the traffic study, the scoping letter and any agency comment letters at the start, in the same request as the site plan and the cost budget. Comment letters are the most informative of the three, because they show what the agency is still unhappy about. We state the pass-by share the engineer used and the capture rate we used, and we show that they reconcile under the identity above. Where they do not, we say which one we believe and why. We carry required mitigation in the project cost build-up as a named line, priced from the engineer's estimate where one exists and from state bid data where it does not, and we do not let it hide inside contingency. We model time to certificate of occupancy with the review rounds in it. And where the access configuration is unresolved, we run coverage at full movement and at restricted movement and report both, so the credit committee sees the range before the DOT decides which end of it the project gets.
For USDA files there is an additional, purely regulatory reason to do this. Roads and local transportation are enumerated factors under technical feasibility in Appendix A, the author must be independent, and the environmental review under Part 1b asks directly whether the project causes a substantial increase in traffic. (12)(18) A study that addresses access in a sentence has not addressed it.
The Midwestern fuel site will get its permit. Most do. The engineer is probably right about the through movement, the figures will be relabeled, the left-turn lane will be designed with whatever storage length the queue analysis supports, and construction will begin some months later than the borrower hoped, at a cost somewhat higher than the first budget showed. Whether the loan behind it was underwritten with those facts or in ignorance of them depends on whether someone read the traffic study as a credit document. It was never written as one. That is the feasibility analyst's job.
Frequently asked questions
Does the SBA require a traffic impact study?
No. SOP 50 10 8.1 does not name or require a traffic impact study, traffic study or traffic impact analysis for 7(a) or 504 loans. The only related instrument in SBA's regulations is the permissive language at 13 CFR 120.160(b), under which SBA may require an appraisal, a survey or a feasibility study. A state DOT or local government may still require a traffic study as a condition of the access permit, and its conclusions will affect the project cost, schedule and projections the lender underwrites.
Does USDA require a traffic study for a B&I or Community Facilities loan?
Not by name. USDA does require a feasibility study in defined cases, including B&I loans above $1 million to a new business, and 7 CFR Part 5001 lists roads, rail and airport infrastructure and the need for local transportation among the technical feasibility factors. Traffic also enters through the environmental review under 7 CFR Part 1b, where a substantial increase in traffic takes a project out of the small-scale development categorical exclusion.
Is a traffic study the same as a feasibility study?
No. A traffic impact study is an engineering report prepared for a road agency to support an access or encroachment permit. It measures the project's effect on the road network and specifies mitigation. A feasibility study is a market and financial report prepared for a credit decision. It measures whether the project can service its debt. They share one input, the traffic count, and they use it for opposite purposes.
What triggers a traffic impact study?
Usually a trip threshold. Many cities and counties use 100 new peak-hour trips. State DOTs vary: Georgia requires a study above 2,000 gross daily trips, North Carolina at 3,000 daily trips, Ohio scales the analysis around 100 peak-hour trip ends, and Texas and Florida leave more to district or local determination. The stricter of the state and local thresholds applies. Gas stations, drive-through restaurants and car washes exceed most thresholds. Self-storage and smaller hotels often do not.
How much does a traffic impact study cost, and who pays for it?
The developer pays. Engineers commonly quote $2,500 to $5,000 per intersection studied, and full studies typically run $10,000 to $30,000 depending on the number of intersections, agencies and analysis scenarios. Agencies often charge review fees on top. The study fee is usually small next to the mitigation it may require, which can run from $30,000 or so for a simple deceleration lane to $500,000 or more for a new signal.
Can turn lanes and other off-site improvements be financed with SBA loan proceeds?
Generally yes, within limits. In the 504 program, land improvements integral to the project are eligible costs unless they are paid through special tax assessments or user fees. In 7(a), site improvements are an eligible use of proceeds, with community improvements such as curbs and sidewalks limited to 5 percent. Whether a particular DOT-required improvement qualifies is a lender or CDC determination on the facts of the deal.
How long does DOT review of a traffic study take?
Plan on months, not weeks. Scoping takes a few weeks, the study itself three to eight, and initial agency review commonly seven weeks or more. Most studies go through at least two rounds of comments and revisions. Four to six months from scoping to an issued permit is common on state routes.
What is the difference between a pass-by trip and a capture rate?
A pass-by trip is a traffic engineering term for a site visit made by a driver already on the adjacent road. A capture rate is a market analysis term for the share of passing vehicles that visit the site. They describe the same vehicles. The capture rate implied by a traffic study equals the pass-by share multiplied by daily site visits, divided by daily traffic on the road, and the two documents in a credit file should agree on it.
September 25, 2026 by Michal Mohelsky, principal of MMCG Invest, LLC, a national SBA and USDA feasibility study consultancy
Request a Feasibility Study https://calendar.app.google/EJzWEz3GCqLY2jU86
Michal Mohelsky, J.D. | Principal | mmcginvest.com
Contact: michal@mmcginvest.com
Phone: (628) 225-1125
Sources
MMCG Invest review of traffic engineer and state DOT district comment-and-response correspondence on a proposed fuel and convenience development, July to August 2026 (project details withheld)
Institute of Transportation Engineers, Trip Generation Manual, 11th Edition, 2021
Transportation Research Board, Highway Capacity Manual, 7th Edition
13 CFR 120.160(b), Loan conditions
7 CFR 5001.3, Definitions, "Feasibility study"
U.S. Small Business Administration, SOP 50 10 8.1, Lender and Development Company Loan Programs, effective October 1, 2026
U.S. Small Business Administration, Information Notice 5000-880695, issuance of SOP 50 10 8.1
13 CFR 120.120(a)(2), Eligible uses of proceeds
13 CFR 120.882, Eligible Project costs for 504 loans
U.S. Small Business Administration, Procedural Notice 5000-872764, 504 construction contingency, September 30, 2025
13 CFR 120.930(c), unused contingency and debenture proceeds
Appendix A to Subpart D of 7 CFR Part 5001, Feasibility Study Components, 85 FR 42494, July 14, 2020
7 CFR 5001.306(a)(3)(i), Specific application requirements for B&I projects
7 CFR 5001.304, Specific application requirements for CF projects
7 CFR 5001.303(b), Applications for loan guarantee
U.S. Department of Agriculture, National Environmental Policy Act final rule, 7 CFR Part 1b, 91 FR 17092, April 3, 2026
USDA Rural Development, stakeholder announcement on implementation of 7 CFR Part 1b, May 12, 2026
7 CFR 1b.4(d)(24), Rural Development categorical exclusions, small-scale site-specific development
USDA Rural Utilities Service, Bulletin 1780-2, Preliminary Engineering Reports for the Water and Waste Disposal Program
7 CFR 5001.307, Specific application requirements for REAP projects
City of Portland, City Code 17.88.050, Transportation Impact Study
Georgia Department of Transportation, Regulations for Driveway and Encroachment Control
North Carolina Department of Transportation, Policy on Street and Driveway Access to North Carolina Highways, and Congestion Management Capacity Analysis Guidelines
Ohio Department of Transportation, State Highway Access Management Manual
Texas Department of Transportation, Access Management Manual, and Traffic and Safety Analysis Procedures Manual, Traffic Impact Analysis chapter
Florida Department of Transportation, Multimodal Transportation Site Impact Handbook
California Senate Bill 743 (2013) and CEQA Guidelines section 15064.3
Institute of Transportation Engineers, Guide to SB 743
Institute of Transportation Engineers, Trip Generation Handbook, 3rd Edition
American Association of State Highway and Transportation Officials, A Policy on Geometric Design of Highways and Streets
National Cooperative Highway Research Program, Report 457, Evaluating Intersection Improvements: An Engineering Study Guide
Federal Highway Administration, Manual on Uniform Traffic Control Devices, Chapter 4C, Traffic Control Signal Needs Studies
Yarger Engineering, published guidance on traffic study cost
Greenlight Traffic Engineering, published guidance on traffic study cost
WGI, published guidance on traffic impact study scope and cost
Portland Bureau of Transportation, transportation impact study review guidance
Texas A&M Transportation Institute, Mobility Investment Priorities, review of TxDOT letting costs for left-turn lane projects
Arizona Department of Transportation, Highway Safety Improvement Program project pricing examples
Washington State Department of Transportation, traffic signal cost guidance
North Carolina Department of Transportation, traffic signal cost guidance
Federal Highway Administration, ITS knowledge resources, signal retiming cost summaries
Florida Department of Transportation, Cost Per Mile Models for Long Range Estimating
Wisconsin Department of Transportation, pass-by trip tables and interim guidance on gas station land uses
Palm Beach County, Florida, Trip Generation Rates table
NACS, State of the Industry data for 2025, released April 2026
National Cooperative Highway Research Program, Report 420, Impacts of Access Management Techniques, 1999
Federal Highway Administration, Safe Access for Everyone Is Good for Business, FHWA-HOP-24-003, August 2024
Iowa State University, Center for Transportation Research and Education, Access Management Research and Awareness Program, business impact findings, 1997
Utah Department of Transportation, Raised Median Economic Impact Study, Report UT-12.17, 2013
Texas Transportation Institute, Report 3904-5, Eisele and Frawley, economic impacts of raised medians on adjacent businesses, 2000
Comments