The Economics of Reinvestment

Correction, 3 November 2026: An earlier version incorrectly treated the December 2025 tariff announcement as implemented from January. The electricity discussion now distinguishes the suspended schedule, the reported July deferral and the separate interim fuel surcharge. Current payable rates remain unverified.

A better question than "expand or refurbish"

Part 1 showed operators committing millions to new stores. The tempting follow-up is that they are neglecting the stores they already have. The evidence does not support that, and does not refute it either, because the comparison needs figures that public accounts do not contain.

The more useful question is where an operator's next dollar earns most: a new store, a refit, replacing refrigeration, maintenance, technology or property. These are not exclusive, and some fund others. The answer depends on whether the operator can measure what each use of capital returns, on the same basis, store by store. This article asks what the public record lets an outsider see, and what an operator would have to measure to see the rest.

We label what we say: a fact is what a document records, a statement is what a company or other source says, a calculation is ours, and interpretation is our reading.

What public accounts reveal: RB Patel as a worked example

RB Patel Group is the one Fiji supermarket operator whose accounts we examined. We use it to show what listed accounts can and cannot tell an outsider. Nothing here implies that it is under-investing, or that it is not.

For the year to 30 June 2026, the audited accounts record:

• Capex. Cash payments of FJ$2.6 million for property, plant and equipment and investment property, of which FJ$1.97 million was investment property. Property, plant and equipment cash payments were therefore about FJ$0.63 million (our subtraction).
• Property, plant and equipment. Additions of FJ$0.65 million, all office equipment and motor vehicles, against a depreciation charge of FJ$1.35 million (in 2025, additions were FJ$1.44 million against depreciation of FJ$1.28 million). The note has no class for plant, fixtures or leasehold improvements.
• Investment property. Eleven properties: two commercial, two residential and seven vacant land.
• Funding. FJ$21.25 million of wholesale corporate bonds issued, and FJ$18.1 million of borrowings repaid. Net, about FJ$3.1 million (our subtraction).

The bond's purpose comes from outside the accounts. The Fiji Times reported, citing the offer announcement, that the money was mainly for a project pipeline in Labasa, Nadawa, Nasinu and Ba, and to repay part of existing bank debt. That is a management statement, and the year's cash flows are consistent with the debt-repayment half.

The line between investing in property and investing in retail operations is visible, but only in part. The accounts separate investment property from property, plant and equipment, and investment property additions were about three times the property, plant and equipment additions (our calculation). But the note has no class that isolates retail equipment, refrigeration or store fit-out: a refit could sit in office equipment, be expensed as repairs, or be capitalised elsewhere. Low additions against depreciation are consistent with under-investment. They are equally consistent with fully depreciated but sound equipment, spending run through other lines, or leased stores fitted out by landlords. Our interpretation: public group accounts distinguish property from operations only coarsely, cannot tell these explanations apart, and cannot reveal the age, condition or return of any individual store.

Evidence status: The FY2026 figures rely on the available financial-statement extraction. Verification against the original PPE note remains outstanding, and a conflicting bond figure is separately documented in the evidence record.

What the large grocers disclose, and what they do not

Coles and Woolworths give a clearer picture of where capital goes, with a warning that their networks, scale and competition are not Fiji's.

In the 2026 financial year, Coles opened 13 supermarkets and completed 71 renewals, according to its ASX results release. Its results presentation, slide 25, splits A$1,189 million of operating capital expenditure into store renewals A$368 million, growth initiatives A$320 million, efficiency initiatives A$201 million and maintenance A$300 million. Woolworths Supermarkets reported 13 net new stores and 67 renewals. On our calculation each renewed about five stores for every one it opened.

Neither discloses what a renewal costs per store or what it returns. Coles says significant projects target returns above its cost of capital (slide 19), without tying that to renewals. It also says its planned acceleration of new stores and renewals is "expected to deliver strong returns consistent with our existing store program" (results release, page 11), a forward-looking statement with no figures.

Our interpretation: even a large grocer that tells investors how much goes into renewals versus growth does not publish the return. If that is the standard at the top of the market, an outsider should expect less from a Fiji operator's accounts, and should not read the absence of store-level returns as evidence about any operator's choices.

What is changing store running costs, and what refit research can say

• Electricity. The December 2025 EFL notice announced a stepped commercial schedule with a stated January 2026 start, including 63.74 cents a unit above 10,000 units a month. FCCC suspended that authorisation. The Fiji Times later reported a deferral to 31 July 2026. Separately, FCCC announced an interim fuel surcharge of 5.91 cents per kWh from 26 May, applied to existing tariffs. These are distinct decisions. As at 3 November 2026, we have not established whether the deferred schedule was subsequently implemented or the interim surcharge changed. The December notice does not establish today’s payable rate. Investment comparisons need the applicable billing schedule and current store bills.
• Refrigerant rules. From 1 August 2025 the Ministry of Environment banned equipment pre-charged with HCFCs. The report does not say whether existing equipment can still be serviced.
• Wages. The listed operator's chairperson says regulated wage rates are up to 87 per cent higher since 2022 (management statement).

US trade press, citing the EPA, puts refrigeration at 40 to 60 per cent of supermarket electricity use, and overseas research shows large swings in case energy from store humidity (Howell 1993: about 5 to 29 per cent between 55 and 35 per cent relative humidity). That is model and laboratory work, mostly from the United States. We found no measurement of equipment age against running cost in Fiji or any tropical supermarket.

On sales, the best evidence we could substantiate is a 2017 field experiment in a US department-store chain. Two remodels cost the same and used identical designs; one lifted sales 12 per cent and the other 1 per cent, which the authors attribute to how large a change shoppers perceived. It covers two stores in a different retail format, so we use it only to show that a refit's sales effect is uncertain. We removed a widely quoted claim of near 50 per cent sales growth after remodelling: it traces no further than a university press summary with no sample, method or retailer.

Sidebar: a cost that equipment comparisons can miss

An international case study, not evidence about Fiji. In October 2026, hackers remotely raised refrigeration temperatures overnight at several hundred French supermarkets, causing food losses and some fires (Reuters, via TimesLIVE). Les Mousquetaires, owner of Intermarché, said about 150 of its stores were held at 50°C for six hours and that it discarded the affected food (same report). France's cybersecurity agency said it was aware of numerous compromises involving remote-control modules made by Danfoss and by Copeland, and the US firm Claroty had reported vulnerabilities in platforms from both in August (AP, via NewsChannel 9). On 2 November, about three weeks later, Reuters reported further temperature spikes at stores in Belgium, Luxembourg, Sweden and Ireland, with an Estonian chain also affected, and said it could not establish who was behind them or whether they were linked to France. Nothing we read establishes a common perpetrator or method.

Statement: Danfoss said the affected units were running outdated firmware, that many were older models that no longer receive updates, and that it had long advised customers to keep supported systems updated and migrate legacy ones (SBS wire report; AFP, via Khaleej Times). That is a supplier's account, we have not tested it, and no source we read establishes how the attackers got in.

Our interpretation: the lesson is not that new equipment is safe or that old equipment is unsafe. A refrigeration comparison that counts energy, repairs, refrigerant and compliance can still leave out who can reach the controls remotely, whether the software is supported and patched, and what one night of lost stock costs. We have not assessed any Fiji operator and make no claim about Fiji. The question worth putting to any operator is who has remote access to its cold chain, and when that was last checked.

Related analysis: Europe’s supermarket fridge cyberattacks: five questions every retailer should ask.

A next-dollar test

Each use of capital pays back in a different place and is overstated in a different way. This table is our interpretation of how an operator could put them on one footing.

The common test is the incremental return on the next dollar, in each store, net of what the same dollar would have earned elsewhere. Only the operator can run it.

What an operator would need to measure

An operator comparing refurbishment with expansion without overstating the return needs at least these store-level records:

1. Asset register by store. Install year, refrigerant type and condition of each major item, so the HCFC rules and failures can be mapped to stores.
2. Store contribution. Sales, gross margin, labour, energy and repairs by store, monthly, on a consistent basis.
3. Project ledger. Every capital project tagged as new site, refit, refrigeration, maintenance, technology or property, with the store it served.
4. Matched control stores. For each refit or opening, comparable stores left alone, so a change can be separated from price, range and competitor moves.
5. Catchment overlap. Which of the operator's own stores share shoppers with a new site, so transfers can be netted.
6. Metered energy. Kilowatt-hours by store, and ideally by refrigeration circuit, to price equipment that is old.

A break-even check keeps the claim honest. The sales uplift a refit needs is its cost divided by the extra contribution that each 1 per cent of sales brings over the horizon the operator accepts. It scales with refit cost and falls with store sales, margin and horizon, so the same refit can be easily recovered in a high-volume store and out of reach in a small one. Setting that required uplift against a wide range of outcomes (the 12 per cent and 1 per cent above show how wide) is more honest than quoting one expected return.

Independent measurement can help only narrowly. Control-store comparisons and energy sub-metering are only as good as the operator's data, and an outside party adds value only where a single operator lacks comparable stores or a neutral baseline. Patternfield has a commercial interest in measurement, and we have written this to stand without it.

What we can and cannot say

Supported:

• Public accounts show where capital goes at group level, and only coarsely. In RB Patel's FY2026 accounts, property additions exceeded property, plant and equipment additions, and the note has no class that isolates retail equipment, refit or refrigeration spending (fact and our calculation).
• Large grocers publish capex splits and renewal counts, not renewal returns (fact).
• Electricity pricing and refrigerant considerations matter to equipment decisions; the December 2025 tariff announcement alone does not establish current payable rates. A refit's sales effect is uncertain (limited research).

Not supported: that any Fiji operator is under-investing; that refits beat new stores or the reverse; that new stores take sales from old ones; any store-level return, for any operator; that Fiji supermarkets are exposed to the kind of attack seen in Europe.

The finding. The strongest defensible conclusion is about evidence, not about capital. Whether to expand or reinvest cannot be judged responsibly from public information. What separates a sound allocation from a hopeful one is whether the operator can measure, store by store, what its next dollar returns. This is a gap in what outsiders can verify, not evidence of inadequate investment or inadequate internal measurement. Even the large grocers' disclosure stops short of returns.