FIFO, LIFO and weighted average

Start with what inventory cost includes, where IAS 2 / Ind AS 2 / ASC 330 are substantively aligned. Cost is purchase price + duties + freight-in + conversion costs (direct labour + allocated production overheads at normal capacity), minus trade discounts. In short, it's everything needed to bring the item to its present location and condition. Abnormal waste, storage between production stages (unless required), selling costs and general admin are excluded and expensed as incurred. In India, GST paid on inputs isn't cost either, because it's a recoverable input credit. Borrowing costs come in only for long-gestation qualifying inventories such as ships or aging whisky, and rarely for trading stock. Techniques like standard cost or the retail method are fine if they approximate actual cost.

Now the cost-flow problem. You've bought identical units at different prices, so which cost goes to COGS and which stays on the balance sheet? Unless items are uniquely identifiable (jewellery, aircraft: specific identification), you have to assume a flow:

• FIFO (first-in, first-out): the oldest costs go to COGS and the newest stay in ending inventory.
• Weighted average (WAC): every unit carries the average cost of the pool.
• LIFO (last-in, first-out): the newest costs go to COGS and the oldest stay in ending inventory. It's US GAAP only. IFRS and Ind AS prohibit LIFO outright. No Indian or IFRS filer can use it.

Worked example: one inventory dataset, three methods

The assumption is about costs. Physics doesn't come into it: a LIFO grocer still sells the oldest milk first. Here's the one dataset you'll run three ways, synthetic Arjun Trading Co. in a rising-price world:

During the year Arjun sells 280 units at ₹180, for revenue of ₹50,400. That leaves 120 units in ending inventory. It's a periodic system, so there's one computation, at year-end.

Take FIFO first. COGS is the first 280 units in: 100 × 100 + 100 × 110 + 80 × 120 = 10,000 + 11,000 + 9,600 = ₹30,600. Ending inventory is the newest 120: 20 × 120 + 100 × 130 = ₹15,400. Check it: 30,600 + 15,400 = 46,000.

Under LIFO, COGS is the last 280 in: 100 × 130 + 100 × 120 + 80 × 110 = 13,000 + 12,000 + 8,800 = ₹33,800. Ending inventory is the oldest 120: 100 × 100 + 20 × 110 = ₹12,200. It passes the same check.

Weighted average gives ₹46,000 ÷ 400 = ₹115/unit. COGS = 280 × 115 = ₹32,200, and ending inventory = 120 × 115 = ₹13,800. That ties too.

What does the comparison tell an analyst?

Now read the table like an analyst. With prices rising, LIFO reports the highest COGS, lowest profit, lowest taxes, and the stalest (lowest) inventory value. FIFO is the mirror image, and WAC falls between them. Neither method changed a single physical fact. Same units, same customers, same cash from sales. The only real economic difference comes from the tax row. LIFO's lower taxable income saves actual cash, here ₹800 (= 25% of the ₹3,200 COGS gap) that a FIFO twin would have paid the tax authority. Everything else is presentation. There are two corollaries. (1) LIFO's P&L is more current, since it sets today's costs against today's revenue and so gives a truer margin signal in inflation, but its balance sheet is ancient. FIFO's balance sheet is current, while its P&L flatters margins with old cheap costs ("inventory profits"). (2) When prices fall, every effect flips.

So why does LIFO exist at all? US tax law. Under the IRS's LIFO conformity rule, a company using LIFO for tax must also use it in its financial statements. It's the only place tax law dictates book accounting. US firms in inflating-cost businesses (groceries, machinery, oil) accept uglier book profits to keep the cash. Kroger, Costco (US merchandise), Deere, Caterpillar, ExxonMobil: each runs LIFO for much or all of its US inventory (check each 10-K's inventory note). India/IFRS filers never face the choice.

What does the LIFO reserve tell you?

The LIFO reserve: the analyst's conversion tool. The SEC requires LIFO users to disclose the LIFO reserve: LIFO reserve = Inventory at FIFO (or current/replacement cost) − Inventory at LIFO. With it you can restate any LIFO company to FIFO and compare it with the rest of the world:

• Inventory_FIFO = Inventory_LIFO + LIFO reserve
• COGS_FIFO = COGS_LIFO − (closing reserve − opening reserve), that is, subtract the change in the reserve
• Pre-tax income_FIFO = reported + Δreserve; Net income_FIFO = reported + Δreserve × (1 − t)
• Equity_FIFO = reported equity + closing reserve × (1 − t) (with a matching deferred-tax liability of reserve × t)
• Cumulative cash taxes actually saved by electing LIFO ≈ closing reserve × t, real money the company kept.

Try it on Arjun Trading. Its reserve is 15,400 − 12,200 = ₹3,200. The opening reserve was zero, because there was one uniform opening layer, so ΔReserve = 3,200, the same as the COGS gap. So the machinery is internally consistent, and you can trust it.

LIFO liquidation: the trap. Suppose a LIFO firm lets inventory quantities fall. It digs into old cheap layers, so decades-old costs flow into COGS, margins spike, and profit gets a one-time, low-quality boost with no operational improvement behind it. You can spot it because the LIFO reserve shrinks alongside falling unit volumes. US filers must also disclose the liquidation's income effect in the inventory note. Be fair, though: a shrinking reserve can also just mean input prices fell, so check volumes before you cry foul.

How does the lower of cost and NRV test work?

Lower of cost and NRV: inventory's impairment test. Inventory is carried at the lower of cost and net realisable value, where NRV = estimated selling price − costs to complete − costs to sell. NRV is an entity-specific net figure. It isn't a market quote. Write-downs go to the P&L, typically within COGS or "changes in inventories". You test item by item or by groups of similar items, and netting the whole warehouse together isn't allowed. Raw materials have a wrinkle. Don't write them down if the finished goods they feed will still sell at or above cost. When a write-down is needed, replacement cost serves as a proxy for NRV.

The regimes split on two points:

Here's the US LCM corridor with numbers. Cost is 100, replacement cost 88, NRV 95, normal profit 10. The ceiling is 95 and the floor 85, so "market" = 88 (inside the corridor). You carry the item at 88.

The reversal asymmetry is easiest to see in a single case. An Indian apparel retailer wrote festival stock down from ₹80 to ₹65, and later NRV recovers to ₹78. It writes the stock back up to ₹78, a reversal of ₹13, capped by the original cost of ₹80. Its US twin stays at ₹65 and simply reports a fat margin when the stock finally sells. Economics are the same on both sides, and only the timing of reported profit differs. Each side also gets a small earnings-management temptation. Under IFRS you write down and reverse later when it's convenient. Under US GAAP you take a "big-bath" now and enjoy juiced margins later. You'll meet the same asymmetry again with long-lived assets, where the same IFRS-reverses/US-never rule reappears.

How can inventory warn you early?

Inventory as an early-warning line (Phase 2 preview). When inventory grows persistently faster than sales, demand has missed the plan. Obsolescence and write-downs follow. Days inventory outstanding, DIO = Inventory ÷ COGS × 365, joins DSO on your standing watch list, along with its trend. Indian filings help you again. Schedule III requires inventory broken into raw materials / WIP / finished goods / stock-in-trade. A rising finished-goods share with flat sales is the classic signature of a write-down on the way.