Portfolio Valuation Methodology Template

Mark Change Explanation, 31 December 2024 to 31 December 2025

Illustrative. Glendevon Instruments is a fictional industrial instruments company, held at a minority stake by a fictional fund. Market approach on a multiple of last twelve months EBITDA. Every figure below was computed with a script and the parts tie to the whole.

The mark rose 16,814,060.00, or 29.5 percent. A valuation memo normally records that as one sentence about a strong year and a firmer market. Neither half of that sentence survives the arithmetic.

Order applied: trading performance, then the three multiple steps, then capital structure, then ownership. Each step holds every later term at its prior value, so the six deltas sum to the change exactly. There is no residual and no line called other. Read as three sentences rather than one: the company earned 14,296,800.00, the market took 4,574,700.00 off, and changing the membership of the comparable set added 7,624,500.00.

The headline is the trap. The selected multiple went from 11.2x to 11.6x, which reads as the sector rerating upward. Held to the five members present at both dates it fell from 11.0x to 10.4x, and every one of the five derated. The sector went down and the mark went up.

Both additions landed in the fourth quarter, and it is the largest quarterly move in the eight quarters the fund has held the position. The additions are 74.0 percent of it. The quarterly figure for the additions is 8,778,120.00 against 9,149,400.00 in the annual bridge, because the annual step is measured with net debt and ownership still at their prior values. That gap of 371,280.00 is why the order gets stated at the top rather than left to the reader.

Comparable Set, by name, at both measurement dates

Nine companies appear across the two dates. Seven were in the set on 31 December 2024 and seven on 31 December 2025. The fund's stated convention is the arithmetic mean of the members present, with no company-specific premium or discount, which is why membership carries the entire judgement in the multiple.

Two rows, and only the second one discloses anything. The count is identical at both dates. The intersection is five of seven, which means two of the seven companies the mark is struck against are different companies from the ones a reader saw last year. The number to report is the size of the intersection, and almost nobody computes it.

The removals are facts. An acquired company stops producing observable market pricing, so both had to go, and the two acquisition multiples were not used because a control transaction is not comparable to a minority holding. The additions are decisions. Both listed months before the measurement date, both price above every surviving member, and together they are worth 9,149,400.00 of the mark change.

The middle row is the judgement no sensitivity table ever carries. Nothing about the membership changes in it. Only the statistic taken across the same seven companies changes, from the mean to the median, and the mark falls 4,389,060.00. A convention chosen after the members are known is an unstated adjustment, so the pack fixes the statistic in the methodology statement and prices the alternative in every memo.

Input Register, aged against the measurement date

Nine of the fourteen inputs behind the Glendevon mark, plus the three exceptions elsewhere in the portfolio. Every row carries three fields a value alone does not: the date it is as of, where it came from, and whether it is observable, an estimate, or a judgement.

Three of the fourteen inputs are past the fund's own ninety day staleness policy and the oldest is 214 days old. Ardgour clears it by two days, which is the kind of margin that only exists as a column. Age is not a defect on its own: the Cairnwell round price is 257 days old and it is an orderly transaction, so its age is the point rather than the problem.

One turn of the multiple is worth 7,315,100.00 to the fund's interest. The two comparables added this period were worth 9,149,400.00, which is more than a turn, so the membership decision is larger than the sensitivity the memo would normally lead with. A reader who knows that treats the set list differently from a reader who does not.

Mark History, eight quarters, every row recomputes

The inputs as they stood at each measurement date rather than a series of marks. Enterprise value is EBITDA times the multiple, equity value is that less net debt, and the mark is the fund's fully diluted share of it. Any row can be rebuilt from the row itself.

The set-members column is the whole reason to keep this sheet. It reads seven, seven, seven, seven, seven, six, five, seven. The six and the five are the two acquisitions, and the return to seven is the two additions, so the membership history is legible from the column even before anybody opens the Comparable Set sheet.

18.2 percent of this book's reported movement came from decisions rather than events. That is a fact rather than a finding: comparable sets have to change. Of the 19 changes, 11 fell in a period where the mark was revised upward and 8 where it was revised downward, and fifteen of the nineteen have no recorded reason. The last number is the one to fix, and it costs nothing to fix going forward.