Priya Raghunathan sells marine hardware out of a warehouse in Annapolis, Maryland. Halyard Supply moves stainless shackles, cleats, turnbuckles, and rigging to boatyards up and down the Chesapeake, and the business has run at a predictable margin for eleven years.
Last spring her bookkeeper migrated the company to new inventory software. Nothing about the business changed. Same suppliers, same customers, same prices on the shelf. But when Priya pulled the first quarterly gross profit report from the new system, the number had moved by several thousand dollars against the same period logic she had used for a decade.
She had not lost a sale or paid more for anything. The new platform defaulted to weighted average costing. Her old one had run FIFO.
Both were correct. Both were accepted by GAAP and by the IRS. They simply answered a different question about which dollars left the warehouse.
The question inventory costing actually answers
You bought the same part at three different prices this year. A customer buys one. Which cost do you send to the income statement?
The physical shackle is identical either way. What differs is the cost flow assumption, which is an accounting choice about which cost moves to cost of goods sold and which cost stays parked on the balance sheet as ending inventory. The assumption does not have to match how stock physically moves through the racks.
That distinction trips up a lot of owners. FIFO does not require you to pick the oldest box off the shelf. It requires you to record the oldest cost.
Cost of goods sold = Beginning inventory + Purchases − Ending inventory
Every dollar of inventory eventually lands in one of two places: this period's COGS, or the closing balance sheet. Your costing method decides the split.
Priya's stainless shackles, run both ways
Halyard bought the same 316-grade shackle three times as steel prices climbed:
| Purchase |
Units |
Unit cost |
Total
cost |
| January opening stock |
600 |
$18.00 |
$10,800 |
| March purchase |
900 |
$21.00 |
$18,900 |
| August purchase |
500 |
$24.00 |
$12,000 |
| Goods available for sale |
2,000 |
|
$41,700 |
She sold 1,500 units at $46 each, so revenue was $69,000. Five hundred units remained on the shelf at year end.
Under FIFO, the oldest costs clear first. All 600 units at $18.00 plus 900 units at $21.00 gives COGS of $29,700. The 500 units left over are valued at the most recent price, $24.00, so ending inventory is $12,000.
Under weighted average, every purchase blends into one cost per unit.
Weighted average cost per unit = Total cost of goods available for sale ÷ Total units available for sale
That works out to $41,700 ÷ 2,000 = $20.85 per unit. COGS becomes 1,500 × $20.85, or $31,275, and the remaining 500 units carry $10,425.
| Line |
FIFO |
Weighted
average |
| Revenue |
$69,000 |
$69,000 |
| Cost of goods sold |
$29,700 |
$31,275 |
| Gross profit |
$39,300 |
$37,725 |
| Gross margin |
57.0% |
54.7% |
| Ending inventory (balance sheet) |
$12,000 |
$10,425 |
The gap on one SKU is $1,575 in reported profit and $1,575 in balance sheet value. Halyard carries 340 active SKUs. Applied across the catalog, the swing on Priya's return ran into the tens of thousands.
Why FIFO looks more profitable when prices rise
Steel does not usually get cheaper. In an inflationary stretch, FIFO pushes your oldest and lowest costs into COGS, which understates current cost, inflates gross margin, and raises taxable income. Weighted average splits the difference by blending old and new, which smooths margin and defers some of the tax.
Across the full life of the product line, both methods run the identical $41,700 through COGS. The disagreement is about timing, and timing is what your tax return and your loan covenant both measure.
Ending inventory deserves its own attention. FIFO leaves the newest, most expensive units on the balance sheet, which inflates current assets and improves your current ratio. For anyone borrowing against inventory on an asset-based line of credit, the method quietly changes the size of the borrowing base.
When you count, it also changes the answer
There is a second decision sitting underneath the first one. A periodic system updates inventory only at period end, so COGS gets backed into by difference after a physical count. A perpetual system records cost on every transaction as it happens.
Under a perpetual system, the weighted average becomes a moving average, recalculated each time new stock arrives:
New average cost = (Existing units × Existing average + New units × New purchase cost) ÷ Total units on hand
Run the same purchases periodically and perpetually, and the totals can differ because a moving average recalculates before some sales rather than after all of them. Neither system is wrong. The number just depends on when you did the math, which is one more reason the method has to be documented rather than assumed.
Specific identification is the fourth option, and it assigns each individual unit its exact acquisition cost. That works for boats, vehicles, jewelry, and serialized machinery. It does not work for a bin of shackles.
Picking on purpose
| Situation |
Better fit |
| Perishable or dated stock |
FIFO |
| High-value, slow-moving, lot-traceable items |
FIFO |
| Interchangeable commodity stock, high turnover |
Weighted average |
| Volatile supplier pricing, many purchases per month |
Weighted average |
| Presenting inventory value to an asset-based lender |
FIFO |
Some owners inherit FIFO because an accountant chose it years ago to make the balance sheet look stronger for a bank, then complain every April about a tax bill they never connected to that decision.
LIFO is the third option, and it is a US-only one. It maximizes COGS while prices climb and minimizes taxable income, but IFRS prohibits it, so any business planning to operate internationally or sell to a public acquirer eventually has to unwind it. The LIFO conformity rule adds another catch: use LIFO on your tax return, and you must use it in your financial statements too.
Whichever you pick, consistency is the requirement. Switching methods is a change in accounting method under the tax code, which means filing Form 3115 and running a section 481(a) adjustment. This is not a dropdown you flip mid-year.
Two other habits keep the numbers honest. Landed cost includes inbound freight, customs duties, and insurance, and shops that omit them understate COGS by 5 to 15 percent. Shrinkage from theft and breakage runs 1 to 3 percent in most warehouses and has to be written off, because pretending the units are still there overstates both assets and profit.
Priya kept the weighted average. Her shackles are interchangeable, they turn over fast, and steady margins make her forecasting easier. What changed was that she made the choice deliberately, instead of letting a software default make it for her.
