Most headcount decisions get argued on feel: culture, commitment, control. The numbers rarely make it onto a slide, and when they do, someone compares a $175,000 salary to an $11,200 monthly invoice and stops there. Below is the full model, with payroll load, vacancy cost, the ramp curve week by week, and the exact hourly rate at which the two options cross.
TL;DR
- A US software engineer costs 1.55x to 1.70x base salary once payroll taxes, benefits, equity, tooling, space and admin are counted. A $175,000 base is a $274,800 annual line.
- The empty seat is the largest cost nobody budgets. Three months to fill a senior role is $68,700 of roadmap that did not ship, plus a $38,500 agency fee on top.
- The ramp curves are nearly identical. The start dates are not. An integrated external engineer starts in 2 to 3 weeks; a US hire starts in 12 to 16.
- Use the parity rate, not the break-even month. Fully loaded annual cost divided by 1,920 hours gives you the hourly rate at which the two options cost the same. For a $175,000 base, that is about $141 per hour.
- At typical LATAM senior rates ($55 to $85), there is no break-even month. The external engineer is cheaper in month 1 and the gap widens every month after.
- Duration only changes the answer between roughly $95 and $141 per hour. Below that band, duration is irrelevant. Above it, the FTE wins immediately.
What a US FTE Actually Costs, Line by Line
Salary is roughly 64% of the total. Here is a senior backend engineer, US remote, at a Series B company:
| Line item | Annual cost | Notes |
|---|---|---|
| Base salary | $175,000 | Senior backend, US remote |
| Target bonus (10%) | $17,500 | Paid at target |
| Employer payroll taxes | $14,400 | FICA, FUTA, SUTA |
| Health, dental, vision (employer share) | $19,500 | Family coverage |
| 401(k) match (4%) | $7,000 | |
| Life and disability insurance | $1,200 | |
| Equity expense | $25,000 | Annual grant value, 4-year vest |
| Software seats and licenses | $4,000 | IDE, repo, cloud dev environment, observability |
| Hardware, amortized | $1,200 | |
| Remote stipend or desk allocation | $6,000 | |
| HR, payroll and benefits administration | $2,000 | |
| Learning budget | $2,000 | |
| Fully loaded annual | $274,800 | 1.57x base |
That is $22,900 per month before you have paid a recruiter. A 22% contingency fee on the base adds a one-time $38,500, due when the person starts.
Two lines move the most between companies. Equity expense swings from $8,000 at a bootstrapped company to $60,000 at a late-stage one, and health coverage swings by family status. Rebuild those two with your own numbers before you argue with anyone about the total.
What an Integrated External Engineer Costs
The external line is short, which is the point. At $70 per hour and 160 hours per month:
| Line item | Annual cost |
|---|---|
| Engineering rate ($70/hr, 160 hrs/month) | $134,400 |
| Employer payroll taxes | $0 |
| Health, retirement, insurance | $0 |
| Equity | $0 |
| Recruiting fee | $0 |
| Hardware | $0 (vendor-provided) |
| Software seats and licenses | $4,000 (same as the FTE) |
| Fully loaded annual | $138,400 |
That is $11,533 per month, and it is the whole number. There is no severance line, no PTO payout, no COBRA subsidy, no unemployment insurance experience rating, and no sunk recruiting fee if the engagement ends at month 9. Exit cost is a 30-day notice period.
The Time-to-Productivity Curve, Week 1 to Week 24
Productivity here means output as a percentage of a fully ramped senior engineer on that codebase. Both people have to learn the same system, so the curves track closely:
| Week | New US FTE | Integrated external engineer |
|---|---|---|
| 1 | 5% | 10% |
| 2 | 15% | 25% |
| 4 | 40% | 45% |
| 8 | 70% | 75% |
| 12 | 85% | 88% |
| 24 | 100% | 92-95% (plateau) |
The external engineer edges ahead early because week 1 is not spent on benefits enrollment, equity paperwork and company orientation, and because vendors pre-screen for the exact stack rather than for general potential. The FTE overtakes around month 6 and keeps climbing on institutional context: who owns what, why the 2022 migration was abandoned, which customer will scream if that endpoint changes.
The five-point gap at the plateau is real and you should price it. It is not worth $136,400 a year.
The Cost of the Empty Chair
A senior backend req in the US takes 60 to 90 days to fill, and the accepted candidate then gives 2 to 4 weeks of notice. Call it 3 months from req approval to first commit, and that is with a functioning pipeline.
Price the empty seat at the loaded cost of the person who is not in it: 3 months x $22,900 = $68,700 of work that did not happen. Your CFO will not see it, because no cash left the building. Your VP of Engineering sees it every sprint, because the roadmap slipped a quarter.
An integrated external engineer starts in 2 to 3 weeks. That difference of roughly 10 weeks is the single largest swing in the whole model, and it lands entirely in year one.
The Worked Example: 24 Months, One Senior Backend Engineer
Assume the req opens at month 0. The FTE starts at month 4. The external engineer starts at month 1. Cumulative cash out:
| Calendar month | US FTE (cumulative) | External at $70/hr (cumulative) | Difference |
|---|---|---|---|
| 3 | $0 | $34,600 | External +$34,600 |
| 6 | $107,200 | $69,200 | FTE +$38,000 |
| 12 | $244,600 | $138,400 | FTE +$106,200 |
| 18 | $382,000 | $207,600 | FTE +$174,400 |
| 24 | $519,400 | $276,800 | FTE +$242,600 |
Now convert to delivered work using the ramp curve. Over 24 calendar months the FTE works 21 months and delivers 19.45 effective engineer-months. The external engineer works 24 months and delivers 21.65 effective engineer-months.
| Model | 24-month cost | Effective engineer-months | Cost per effective month |
|---|---|---|---|
| US FTE | $519,400 | 19.45 | $26,704 |
| External at $70/hr | $276,800 | 21.65 | $12,785 |
The FTE costs 2.09x per unit of shipped work over the first two years, and delivers 11% less of it.
The Parity Rate: Where the Two Lines Actually Cross
Stop asking for a break-even month and calculate the parity rate instead:
Parity rate = (fully loaded annual cost minus shared tooling) / 1,920 hours
| US base salary | Fully loaded annual | Multiplier | Parity hourly rate |
|---|---|---|---|
| $130,000 | $205,700 | 1.58x | $105 |
| $145,000 | $228,800 | 1.58x | $117 |
| $175,000 | $274,800 | 1.57x | $141 |
| $210,000 | $355,800 | 1.69x | $183 |
Below the parity rate, the external engineer is cheaper on a run-rate basis and stays cheaper forever. Above it, the FTE is cheaper from the first month of employment. The only zone where duration decides anything is between the rate that offsets the recruiting fee and vacancy head start (about $95 per hour against a $175,000 base) and the parity rate itself.
How Expected Duration Changes the Equation
Here is the month at which the external engineer becomes cumulatively cheaper than the $175,000 FTE, by rate:
| External rate | Monthly cost | External becomes cheaper at |
|---|---|---|
| $70 | $11,533 | Month 1 (never behind) |
| $95 | $15,533 | Month 5 |
| $110 | $17,933 | Month 7 |
| $120 | $19,533 | Month 9 |
| $130 | $21,133 | Month 18 |
| $135 | $21,933 | Month 32 |
| $141+ | $22,900+ | Never |
This inverts the usual assumption that contractors are fine short-term but employees win over the long run. At LATAM senior rates, longer duration favors the external engineer, because the monthly gap of $11,367 compounds while the one-time $38,500 recruiting fee does not.
Two things push the FTE case back the other way over a 3-year-plus horizon. US merit increases run 4% to 6% a year with equity refreshes on top, while external rates typically move 3% to 5%, so the gap widens rather than closes. But if the FTE stays five years, the domain depth that shows up as the 5-point plateau gap starts compounding into architectural decisions that are hard to price. If the role is a permanent owner of a system that will outlive three product cycles, hire the FTE and stop modeling. If the role is capacity against a roadmap, the model above is the whole answer.
Frequently Asked Questions
What multiplier should I use on base salary?
Between 1.55x and 1.70x for a US software engineer, with the spread driven almost entirely by equity expense and family health coverage. Build it from actual line items once, then reuse the multiplier for planning. Anything below 1.4x means you left out equity, space, or benefits admin.
Should the vacancy cost go in the comparison?
Not in the cash comparison, because no money leaves the company while the seat is empty. Put it in the delivery comparison, where it belongs. Three months of a $22,900 monthly loaded cost is $68,700 of work not done, and that is what makes the 10-week difference in start date matter.
Does an external engineer really ramp as fast as a new hire?
Through week 12, yes, within a few percentage points, because both are learning the same unfamiliar codebase. The difference appears after month 6, where a full-time employee keeps accumulating institutional context and the external engineer tends to plateau around 92% to 95% on domain-heavy work.
At what point is hiring an FTE actually cheaper?
When the external rate exceeds the parity rate, which is fully loaded annual cost divided by 1,920 hours. For a $175,000 base that is $141 per hour, well above typical LATAM senior rates and inside the range for US onshore contractors. Below $95 per hour, the FTE is never cheaper at any duration.
How do I handle the equity line at a private company?
Use the annual grant value your board approved at the last 409A price, divided by the vesting period. If you are pre-revenue and the number feels unreal, run the model twice, once with equity and once without, and show both. The conclusion rarely changes because equity is 9% of the total in the example above.
What breaks this model?
Three things. On-call and incident response, if the external engagement excludes it and you have to backfill coverage internally. Compliance requirements that force US-based employment for specific systems. And engagements shorter than four months, where nobody clears the ramp curve and the cost per effective engineer-month roughly doubles for both options.
