Choose a depreciation convention that holds up in a heavy fourth quarter.

The dropdown every fixed asset register reaches.

Setting up a fixed asset register, or cleaning one up before a first audit, eventually reaches a dropdown: full month, mid month, mid quarter, half year. Search for which one to pick and the results are mostly software help pages defining each option, plus a few explainers on the half year convention. The definitions are accurate. What they leave out is which one belongs on the books, and what the choice costs in a year when purchases do not arrive evenly.

The useful starting point is that conventions never change the total. An asset costing 150,000 dollars depreciates 150,000 dollars over its life under every one of them. They only move expense between years, and the year an auditor tests is a single year. So the choice is really about how close the first year lands to the truth, and that depends on when in the year you buy.

What the book standards actually ask for.

Neither IFRS nor US GAAP names a convention. Reading IAS 16 in the consolidated text, paragraph 50 asks for the depreciable amount to be allocated on a systematic basis over the useful life, and paragraph 55 says depreciation begins when the asset is available for use, meaning in the location and condition necessary for it to operate as management intends. US GAAP in ASC 360 asks for the same thing in different words, a systematic and rational allocation over the useful life.

Read literally, the standard gives you a start date. A convention is a shortcut that approximates the start date, and the reason it is acceptable sits in IAS 8 paragraph 8: IFRS policies need not be applied when the effect is immaterial, though it is inappropriate to leave immaterial departures in place to achieve a particular presentation. So the honest test for a book convention is simple. Does it land close enough to depreciation based on time in service that nobody reading the accounts would care about the difference?

IAS 16 paragraph 61 adds one more requirement: the depreciation method is reviewed at least at each financial year end. A convention is part of how the method is applied, so a dated note that you looked at it is cheap evidence to keep.

Where the half year habit comes from.

The conventions most people know are tax rules. Reading IRS Publication 946 directly, MACRS uses mid month for real property, mid quarter when the depreciable bases placed in service in the last three months of the tax year are more than 40 percent of the year’s total, and half year otherwise. Treasury Regulation 1.168(d)-1 states the same 40 percent test in the regulation itself.

Those rules are set by statute and they answer a different question: how much deduction a taxpayer may claim. They are also less central than they used to be. Publication 946 notes that the 100 percent special depreciation allowance was reinstated for certain qualified property acquired and placed in service after 19 January 2025, unless you elect out. Where it applies, there is little left for a tax convention to spread. The books still need a systematic allocation over the useful life, so the book convention is a separate decision, and it is worth making on its own terms. If you run separate book and tax schedules, our page on book and tax depreciation books covers how the two sit side by side.

One uneven year, six answers.

Here is an example year for a made up software company. It buys 30,000 dollars of laptops in February, 60,000 dollars of servers in June, and a 150,000 dollar network and server refresh that goes live on 18 November. Everything is depreciated straight line over 36 months with no residual value, and the financial year is the calendar year. November carries 62.5 percent of the year’s spend, which is the pattern you get when budget is released late or a refresh is timed to year end.

The baseline is depreciation measured by the days each asset was actually in service, which is what paragraph 55 describes. Against it we ran the textbook versions of five conventions: full month counts the month of purchase as a whole month, next month starts the month after, mid month gives the first month half, mid quarter treats each asset as placed in service at the middle of its calendar quarter, and half year gives every asset six months in its first year.

Mid month lands 216.89 dollars from the baseline, within 1 percent. Full month and next month miss by roughly half a month per asset in opposite directions, 3,550 dollars over and 3,116 dollars under. Mid quarter is 1,883.56 over, mostly because the June servers are treated as arriving in mid May.

Half year books 40,000 dollars against 25,616, an overstatement of 14,383.56 dollars, or 56 percent more depreciation than time in service supports. Look at the rows and it is wrong in both directions at once. The February laptops get 5,000 dollars against 8,904 earned, and the November refresh gets 25,000 against 6,027. Half year only works when those two errors cancel, which needs purchases spread evenly across the year.

The tax code has noticed the same thing. A year with more than 40 percent of its basis in the final quarter is exactly the year MACRS stops allowing half year. This example is at 62.5 percent. If the statute will not accept half year in this pattern, it is a weak approximation for the books in the same year.

The date matters more than the convention.

Now change one fact. The network refresh is invoiced on 18 November but racked and switched on 6 January. Under paragraph 55 it was not available for use in 2026, so its book depreciation for the year is zero. A register that runs from the invoice date books 6,250 dollars under mid month or 8,333 under full month, and no choice of convention fixes that.

That is the more common finding in a first audit. Cut off testing on additions near the year end asks when each asset started working, and the answer has to come from an installation record, a commissioning email or a delivery note, not from the bill. A register that holds the in service date as its own field, separate from the invoice date, answers it in a minute.

How this works in Rivane.

In Rivane each fixed asset carries its own depreciation convention: full month, which is the default, mid month, mid quarter or half year. A value outside those four is rejected when the asset is saved rather than quietly replaced with the default, because a silent fallback produces a schedule that looks fine and is wrong at both ends.

The engine applies a convention as a half rate window starting in the month depreciation begins: one month for mid month, three months for mid quarter, twelve months for half year. The amount held back is recognised at the end of the life, so the schedule runs one, two or six months longer, and the final period absorbs any rounding so the total equals cost less residual to the cent. For mid month that matches the textbook column above exactly. For mid quarter and half year the window follows the asset’s own start month rather than your calendar quarter or financial year, which is one more reason we would point a book register at mid month.

Depreciation runs from the asset’s acquisition date, and the vendor bill the asset was capitalised from is stored as a separate link. So equipment invoiced in November and installed in January can carry a January date and still point back to its bill. Straight line and double declining balance are supported, with double declining switching to straight line when that gives the larger charge. On disposal, any depreciation not yet run is posted through the month of disposal before the gain or loss is calculated. Rivane keeps one depreciation book per asset, so a tax schedule with its own convention lives with your tax provision rather than in the register.

# A fixed asset record that can explain its own first year.
# Shape only. No real company, vendor or asset.

fixed_asset:
  name:                 Network and server refresh
  category:             Computer equipment
  cost:                 150000.00
  residual_value:       0.00
  useful_life_months:   36
  method:               STRAIGHT_LINE
  convention:           MID_MONTH         # stored on the asset, from policy
  invoice_date:         2026-11-18        # what the vendor billed
  in_service_date:      2026-11-18        # what depreciation runs from
  source_bill:          BILL-2026-0412    # link to the supporting document

schedule (MID_MONTH):
  2026-11   2,083.33    # half month in the first period
  2026-12   4,166.67    # full months from here
  ...       4,166.67    # 35 full months in total
  2029-11   2,083.22    # the deferred half; the final period absorbs rounding
  total   150,000.00    # always cost less residual, whatever the convention

# If the kit was invoiced on 18 November but racked on 6 January,
# in_service_date is 2027-01-06 and the 2026 charge is zero.

A checklist for the register.

• Write the book convention into the fixed asset policy as a decision, with one sentence on why it approximates time in service. An auditor reading the policy should not have to infer the convention from the schedule.
• Pick mid month unless you have a reason not to. In the worked year it lands within 1 percent of the days in service figure, and in a system it costs nothing extra to run.
• Keep the tax convention out of the book decision. MACRS conventions are set by statute and can change year to year with the 40 percent test, while the book convention should stay put.
• Store the in service date on every asset, separately from the invoice date, and make it the date depreciation runs from. This is the field that moves the number most.
• Look at your fourth quarter share of additions each year. If it runs well above a quarter of the year, check the gap between your convention and days in service before the auditor does.
• Run the comparison in the worked example on your own additions once a year: convention against days in service, by asset and in total. It takes an hour with an export and gives you the evidence that the approximation is immaterial.
• Apply the convention at both ends of the life. If acquisitions get a half month, decide and document what disposals get, and confirm the system does the same.
• Review the method and convention at each year end and record that you did. IAS 16 paragraph 61 asks for the method review explicitly, and a dated note is the evidence.

Where this usually goes wrong.

• Carrying the half year convention onto the books because the tax return uses it. It only matches time in service when purchases are spread evenly through the year, which is exactly the year that never seems to happen.
• Starting depreciation on the invoice date. Kit invoiced in November and racked in January has no book depreciation in the old year under IAS 16 paragraph 55, whatever the convention says.
• Changing the convention in the system without a policy note. A schedule that silently switches from full month to mid month on some assets reads, to a reviewer, like an error in the ones that changed.
• Mixing conventions inside one asset class without a reason. Two identical laptops depreciating differently is the first thing a sample test finds.
• Assuming the total over the life will forgive the timing. It does in the end, since every convention depreciates the same total, but an audit looks at one year, and the timing is the whole question for that year.
• Using a 100 percent tax write off as a reason to stop caring about the book schedule. The special allowance removes the convention from the tax return for qualifying property. The books still need a systematic allocation over the useful life.

Data and system considerations.

• Store the in service date and the acquisition or invoice date as two fields. They agree most of the time, and the times they differ are the ones that matter at a year end.
• Store the convention on the asset rather than only on the asset class. Defaults belong on the class; the value actually used belongs on the record, so a later change of default does not rewrite history.
• Keep the link from the asset to the vendor bill it was capitalised from. It is the fastest route from a sample selection to the supporting document.
• Reject an unknown convention value at entry rather than defaulting it. A silent fallback to full month produces a schedule that looks fine and is wrong at both ends.
• Make sure the schedule closes to cost less residual exactly, with the last period absorbing rounding. A register that leaves a few cents on a fully depreciated asset generates reconciling items forever.
• On disposal, post any depreciation not yet run through the disposal month before computing gain or loss, so the gain or loss is not quietly absorbing missed depreciation.

What the audit evidence looks like.

• The fixed asset policy, showing the book convention, the reason it approximates time in service, and the date it was last reviewed.
• For each addition in the year: cost, in service date, invoice date, convention, life and first year depreciation, with the source bill attached or linked.
• The annual comparison of convention against days in service, by asset and in total, showing the difference is immaterial.
• For additions near the year end, evidence of the in service date itself: an installation record, a commissioning email, a delivery note. This is what a cut off test asks for.
• A rollforward from opening cost and accumulated depreciation to closing, tying to the general ledger, with additions, disposals and the depreciation charge as the movements.

Questions worth asking internally.

• Which convention does our register use today, and did someone choose it or did the software default to it?
• Is our book convention the same as our tax convention, and if so, is that a decision or an accident?
• What share of last year’s additions landed in the final quarter?
• Does depreciation in our system start from the invoice date or from the day the asset went into use?
• If an auditor picked our three largest additions in December, could we show when each one actually started working?

Write the convention down and check it once a year.

The good news is that this is a small decision with a clear answer. Record when each asset actually went into use, choose mid month unless you have a reason not to, write it into the policy, and once a year compare it against days in service on your own additions. That comparison is the evidence that the shortcut is immaterial, and in most years it takes an hour.

Every convention starts from the same annual straight line or declining balance charge. If you want that figure for an asset before deciding how to treat its first year, the Rivane depreciation calculator works it out from cost, residual value and useful life.

Sources

• IAS 16 Property, Plant and Equipment, as consolidated in Commission Regulation (EU) 2023/1803 on EUR-Lex. Read in the text for paragraph 50 (the depreciable amount allocated on a systematic basis over the useful life), paragraph 55 (depreciation begins when the asset is available for use, meaning in the location and condition necessary for it to operate as management intends), and paragraphs 60 and 61 (the method reflects the expected pattern of consumption and is reviewed at least at each financial year end, with a change treated as a change in accounting estimate under IAS 8).
• IAS 8, paragraph 8, in the same consolidated text. Source for the statement that IFRS accounting policies need not be applied when the effect of applying them is immaterial, and that it is inappropriate to make or leave uncorrected immaterial departures to achieve a particular presentation. This is the paragraph that makes a convention an acceptable approximation rather than a departure.
• IRS Publication 946 (2025), How To Depreciate Property. The page was downloaded and read directly. Source for the three MACRS conventions and when each applies: mid month for nonresidential real property, residential rental property and railroad grading or tunnel bore; mid quarter when the depreciable bases placed in service in the last three months of the tax year exceed 40 percent of the year’s total; half year otherwise. Also the source for the 100 percent special depreciation allowance reinstated for certain qualified property acquired and placed in service after 19 January 2025, unless the taxpayer elects out.
• Treasury Regulation 26 CFR 1.168(d)-1, Applicable conventions, half year and mid quarter conventions, on eCFR. Read in the text for the 40 percent test as the regulation states it: the mid quarter convention applies if the aggregate basis of property placed in service during the last three months of the taxable year exceeds 40 percent of the aggregate basis of property placed in service during the year.
• Corporate Finance Institute, Half year convention for depreciation. Read as an example of the answers a search currently returns: a clear definition of the convention, with no guidance on choosing one for the books or on what it costs in an uneven year.
• Rivane fixed assets engine, described from the product as built rather than from marketing copy, and the Rivane depreciation calculator linked here. The per asset depreciation convention, the half rate acquisition window with the deferred amount recognised at the end of the life, the final period absorbing rounding so the total equals cost less residual, the rejection of an unknown convention, the separate source bill link, and the catch up depreciation posted on disposal are described as they are built in the product today. Rivane keeps one depreciation book per asset; a separate tax book is not part of the product.

A note on how this was researched. IAS 16 and IAS 8 were read in the consolidated text of Commission Regulation (EU) 2023/1803, and IRS Publication 946 and Treasury Regulation 1.168(d)-1 were read as the documents themselves rather than through a summary. Paragraph numbers are quoted only from text we read. The worked year is an example with placeholder assets and dates, and every figure in the table is our own calculation, which you can repeat with a spreadsheet. Nothing here is tax advice; the MACRS rules are described only to separate them from the book decision.

On embedded social posts: we looked for public posts from standard setters, auditors or finance practitioners discussing book depreciation conventions and found none we could verify as current, relevant and authentic, so this article uses ordinary citations instead.