Calendarization

(c) Fiscal-year misalignment: calendarize. India runs April to March. US filers scatter: Apple ends late September, Microsoft June, Walmart January and Costco early September. Many retailers use 52/53-week years, and a 53rd week adds ~1.9% of "growth" that the 10-K discloses and you must strip. "FY24 vs FY2024" can misalign by up to a year of macro, and that's fatal across a break like 2020 or the 2021-22 inflation spike. Here are the fixes, cheapest adequate one first. (1) If both fiscal years end within ~3 months of each other and the period was calm, compare as-is and footnote the offset. (2) Calendarize. Rebuild both companies to a common 12 months from quarterly or half-yearly data. For a June-ending filer, calendar-2025 = H2 of its FY25 (Jan to Jun 25) + H1 of its FY26 (Jul to Dec 25). (3) For ratios, use TTM (trailing-twelve-month) figures ending at the nearest common quarter. WE2 works it fully. India has its own live example. Nestlé India moved from a December to a March fiscal year, and its transition period ending March 2025 ran fifteen months. Any growth or margin comparison that uses that row raw is broken. Annualize it, and expect screener rows to get it wrong until the databases catch up.

The LTM build, worked three ways

Calendarisation is a rule until you've built one. Then it's a reflex. Work the three cases that cover almost every misalignment you'll meet.

Case one: a stub quarter on top of an annual report. Cordillera Instruments (synthetic, US, December fiscal year, $m) filed a 10-K for FY2025 showing revenue of 4,800 and EBITDA of 960. Its Q1 FY2026 10-Q shows revenue 1,290 and EBITDA 258, against Q1 FY2025 comparatives of 1,140 and 222 printed in the same document. The trailing twelve months to 31 March 2026 is the annual figure, plus the new quarter, minus the quarter it replaces:

LTM revenue = 4,800 + 1,290 − 1,140 = 4,950
LTM EBITDA  =   960 +   258 −   222 =   996

The margin moves from 20.00% to 20.12%, and both numbers came out of two filings and one subtraction. Every 10-Q carries its own prior-year comparative, and that's what makes the subtraction free. You never need the earlier year's filing.

Case two: two stub quarters. Add the Q2 FY2026 10-Q, with revenue 1,340 and EBITDA 272 against comparatives of 1,205 and 238. Now the half-year replaces the half-year:

LTM to 30-Jun-2026 revenue = 4,800 + (1,290 + 1,340) − (1,140 + 1,205) = 5,085
LTM to 30-Jun-2026 EBITDA  =   960 + (  258 +   272) − (  222 +   238) = 1,030

That's a margin of 20.26%, with revenue 5.94% above the FY2025 figure. Compare it with the shortcut a hurried analyst reaches for, doubling the half-year: 2 × 2,630 = 5,260. That overstates LTM revenue by 175, and it would misprice the company by the same 3.4%. Doubling a half assumes the two halves are identical, and no business obliges.

What if a company changes its fiscal year end?

Case three: the fiscal-year-end change and its stub period. Suppose a company moves from a September year-end to a March year-end and reports a six-month transition period, October 2024 to March 2025, with revenue 2,520 and EBITDA 470. Doubling that stub gives 5,040 of "annual" revenue. The real twelve months to 31 March 2025 is the transition stub plus the April-to-September 2024 half from the last September-year accounts, with revenue 2,280 and EBITDA 425:

12m to 31-Mar-2025 revenue = 2,280 + 2,520 = 4,800     EBITDA = 425 + 470 = 895   (margin 18.65%)

The naive annualisation overstates revenue by 240, about 5%, because the stub caught the seasonally stronger half. India's live case is Nestlé India's move from December to March, whose transition period ran fifteen months. The arithmetic is the same, and the error points whichever way seasonality points. Identify the stub, name its months, and rebuild twelve of them.

How do you adjust for a 53rd week?

The 53rd week. A US retailer on a 52/53-week calendar reports FY2025 revenue of 10,600 against FY2024's 10,000, growth of 6.0%. But FY2025 had 53 weeks. Scale the 53-week year back to a comparable 52: 10,600 × 52 ÷ 53 = 10,400. Clean growth is 4.0%, and a full 2.0 points of the headline was the calendar. One extra week is worth 53 ÷ 52 − 1 = 1.92% of a normal year, and it inflates the following year's comparison in the opposite direction when the extra week drops out. The 10-K states the week count. Find it before you quote a growth rate for any retailer, restaurant chain or distributor.