The strategic use of debt: a Threefold Capital perspective

Next update. The Reserve Bank's next decision on the cash rate is due on 3 November 2026. We will update the rates and figures in this article after it.

Two buyers each pay $2 million for the same business, one that produces $200,000 a year of cash before interest. The first pays with its owners' money. The second has its owners put in $300,000 and borrows the rest from a bank, a specialist lender and the person selling the business. If the business is sold five years later for what it cost, the first buyer's owners have earned 10.0% a year and the second's 21.3%. If it is sold for 30% less, the first buyer's owners still earn 4.5% a year. The second buyer's owners receive nothing from the sale, and the seller, who agreed to be repaid after the other two lenders, loses all $300,000 of its loan.

Nothing about the business differs between the two cases, only the way it was paid for. That alone decides who gains, who loses, and in what order.

This is the last of seven articles on how lending works in Australia, and the first to sit on the borrower's side of the table. It is written from the borrower's side, and its last section sets out the principles we hold, as views.

Fall in value that wipes out the owners at the sale when 85% of a $2 million purchase is borrowed

15%

Threefold Capital model

Extra money the owners must put in over five years to keep a loan of 70% of the price current, in the worst case

$560,000

Threefold Capital model

The least notice a bank that has adopted the Banking Code must give a small business not in default that it will not extend a loan not fully repaid by instalments

3 months

Banking Code of Practice

When debt adds value, and when it destroys it

A lender's return is fixed in advance. It receives its interest and its money back, and nothing more however well the business does, unless it has bargained for a share of the gain, as some mezzanine lenders do. The owners receive whatever is left. That difference is the whole of the arithmetic of leverage.

Our first model shows it with a hypothetical $2 million purchase of an established business, such as a trade distributor or an engineering firm, which produces $200,000 of cash a year before interest and is sold after five years for what it cost. Without debt, the owners earn 10.0% a year. If they borrow $900,000, 45% of the cost, from a bank at 7.5% interest only, the interest is $67,500 a year. The owners keep $132,500 a year on the $1,100,000 they put in, and their return rises to 12.0%. The business earns 2.5 percentage points more than the bank charges, and on $900,000 of borrowed money that is $22,500 a year, which adds about two points to the return on the owners' money. The rule behind the figures: the owners' return is the return on the whole investment, plus the gap between that return and the interest rate for every dollar borrowed per dollar of their own.

All of this is before tax. Interest on money borrowed to earn business income is generally deductible to the borrower, which lowers the cost of debt by the tax saved, while a dividend is not; dividend imputation, which passes company tax on to shareholders as franking credits, narrows that advantage at the shareholder level. Our models leave tax out, and the second takes its cash before interest, so it does not count the saving and overstates what the loan takes.

The table runs the same purchase at four levels of cash flow, from $100,000 a year to $250,000, and four levels of bank debt.

Illustrative model

Return on the owners' money a year, by the investment's cash flow and the share of the cost borrowed at 7.5%

	Bank debt, share of the cost
Cash flow a year	0%	30%	50%	70%
$100,000	5.0%	3.9%	2.5%	−0.8%
$150,000	7.5%	7.5%	7.5%	7.5%
$200,000	10.0%	11.1%	12.5%	15.8%
$250,000	12.5%	14.6%	17.5%	24.2%

The purchase costs $2 million, so cash flow of $100,000 a year is a return of 5% on it, $150,000 is 7.5%, the rate on the debt, $200,000 is 10% and $250,000 is 12.5%. The debt pays interest only, and the business is sold after five years for what it cost. Where the interest is more than the cash flow, the owners are assumed to pay the difference each year. Losses carry a minus sign and are shown in red.

Debt also destroys value in ways a calculation of returns does not show. It can force a sale at the wrong time, because a lender owed money on an asset worth less wants to be repaid, not to wait for a recovery. It takes cash the business may need for stock, staff or equipment, because loan payments come first. It can move control, because a breach of the loan's terms gives the lender rights over decisions the owners used to make alone. And it costs most when it is needed most: a business that must refinance in a bad year does so on the terms of a bad year, as Part 3 describes. None of this appears in the table above, and we think these costs are the larger part of the decision.

Avoiding debt is no guarantee of safety either. The Reserve Bank notes that around 70% of the companies that entered insolvency in 2024-25 owed nothing to secured creditors, and around 75% had fewer than 20 full-time employees.⁶ Businesses fail from losses, unpaid tax and unpaid suppliers as well as from bank loans. What borrowing changes is how much room a business has before it reaches that point, and who decides what happens when it does.

The capital structure: who is paid first, and who loses first

A capital structure, often called a capital stack, lists everyone who has put money into an asset in the order they are paid. Money at the bottom of the stack is paid first and lost last; money at the top is paid last and lost first. Each layer's price reflects where it sits: a lender at the bottom accepts a lower rate because others will absorb losses before it does, and the owners at the top accept the first loss in exchange for everything left once the others are paid.

Threefold Capital interpretation

The layers of a capital structure, from the first paid to the last

Layer	Usually provided by	How it is paid	If things go wrong
Senior debt	A bank or other lender	Interest, often with scheduled repayments of principal. Paid first	Can enforce its security after the notice its agreement requires. A bank that has adopted the Banking Code gives a small business at least 30 days' notice of a missed payment before demanding repayment, with some exceptions
Mezzanine debt	A private credit fund or specialist lender	A higher rate, sometimes added to the loan rather than paid, and sometimes a share of any gain. Paid after senior debt	Usually limited by its deed with the senior lender until that lender is repaid or agrees
Vendor finance	The seller of the business	Part of the price, paid later, with or without interest. Usually ranks behind the bank under a deed	Can pursue what it is owed only as far as its deed with the bank allows
Preferred equity	Investors	A set return before the ordinary shareholders, and its money back before them on a sale. Ranks behind every lender	Cannot enforce security. Its rights are those in the company's constitution or the terms of issue, and a dividend may be paid only when the company's assets exceed its liabilities by enough and its creditors are not prejudiced
Ordinary equity	The owners	Whatever is left after everyone else	Controls the board and the strategy, until a lender's rights intervene

A general description. The terms of any real layer are set by its own agreement, and the order between lenders by the security each holds and any deed they sign.

Our first model puts these layers together for the same $2 million purchase. The bank lends at 7.5%, the mezzanine lender at 14% and the vendor at 6%, all interest only and repaid from the sale. These rates are our assumptions, not quotes. For context, the cash rate target is 4.60%,⁵ so the bank's 7.5% sits 2.9 percentage points above it, between the 7.29% that the model in Part 1 builds up from the cash rate for a secured commercial property loan and its 9.39% for an unsecured small business loan. The stretched structure is also one that, in our reading of lenders' published terms, a bank would be unlikely to write: senior debt against a business usually amortises, as the next section shows, and a loan with mezzanine debt behind it would normally bar the owners from taking cash out while cover was as thin as 1.47×. It is in the model to show what happens when the layers are piled high.

Illustrative model

Model: build your capital structure

Each scenario pays for the same $2 million purchase in a different way. Choose one, or set the layers and their rates yourself, then change the value at the sale to see who is repaid.

All equityBank debt, 45%Bank and vendor financeBank, mezzanine and vendorStretched, value down 30%

Without JavaScript the model shows one scenario: Bank debt, 45%.

The investment

Total cost of the investment

$2,000,000

−+

Cash flow before interest, a year

$200,000

−+

Years until it is sold

5 years

−+

Change in value at sale

0%

−+

The capital structure

Senior bank debt, share of cost

45%

−+

Mezzanine debt, share of cost

0%

−+

Vendor finance, share of cost

0%

−+

The cost of each layer

Senior debt interest rate

7.50%

−+

Mezzanine interest rate

14.00%

−+

Vendor finance interest rate

6.00%

−+

Return on the owners' money 12.0%debt 45%

Return on the owners' money, a year

12.0%

Return on the whole investment, a year

10.0%

Cash flow, times the interest

2.96×

Fall in value that wipes out the owners

55%

At a sale for $2,000,000, the bank is repaid first, and the owners receive $1,100,000 for the $1,100,000 they put in.

The owners' money takes the first loss: a fall in value of 55% would wipe it out at sale. Only a larger fall would cost the bank.

Borrowing 45% of the cost turns a return of 10.0% a year on the whole investment into 12.0% on the owners' money.

Build your capital structure

Illustrative model

The same purchase, financed five ways

Line	All equity	Bank debt, 45%	Bank and vendor finance	Bank, mezzanine and vendor	Stretched, value down 30%
Senior bank debt	$0	$900,000	$900,000	$1,200,000	$1,200,000
Mezzanine debt	$0	$0	$0	$200,000	$200,000
Vendor finance	$0	$0	$300,000	$300,000	$300,000
The owners' money	$2,000,000	$1,100,000	$800,000	$300,000	$300,000
Interest a year	$0	$67,500	$85,500	$136,000	$136,000
Cash flow, times the interest	n/a	2.96×	2.34×	1.47×	1.47×
Cash to the owners a year	+$200,000	+$132,500	+$114,500	+$64,000	+$64,000
Value at sale	$2,000,000	$2,000,000	$2,000,000	$2,000,000	$1,400,000
Repaid to the bank	$0	$900,000	$900,000	$1,200,000	$1,200,000
Repaid to the mezzanine lender	$0	$0	$0	$200,000	$200,000
Repaid to the vendor	$0	$0	$300,000	$300,000	$0
To the owners	$2,000,000	$1,100,000	$800,000	$300,000	$0
Return on the owners' money, a year	10.0%	12.0%	14.3%	21.3%	2.2%
Return on the whole investment, a year	10.0%	10.0%	10.0%	10.0%	4.5%
Owners' money back, times	1.50×	1.60×	1.72×	2.07×	1.07×
Fall in value that wipes out the owners	100%	55%	40%	15%	15%

Cash flow before interest is $200,000 a year in every scenario, and the business is sold after five years. The last scenario is the stretched structure sold for 30% less than it cost. Amounts are rounded to the nearest $1,000, and interest and cash flow to the nearest $100.

The model has no preferred equity. Had the stretched structure included some, it would sit between the vendor and the owners: paid its set return before them and repaid before them on a sale, but behind every lender, so in the 30% fall it too would have received nothing. For the company, preferred equity is easier to live with than mezzanine debt. A missed dividend is not a default, and the law lets a company pay a dividend only when its assets exceed its liabilities by enough, the payment is fair to shareholders as a whole and its creditors are not materially prejudiced (Corporations Act 2001, section 254T),⁸ a test a lender's interest never waits for. For the investor who holds it, that is exactly its weakness. We believe it should therefore cost the company more than mezzanine debt and less than ordinary equity, compared after tax, since interest is generally deductible and a dividend is not, and that its terms deserve as close a reading as a loan agreement.

From returns to cash: repayments and covenants

The first model lets the debt wait until the sale. Many business loans do not: they require principal to be repaid in instalments over the loan's life, called amortisation, and lenders' published terms tie how much may be borrowed to how soon it is repaid. Macquarie's commercial lending guide for brokers, dated May 2026, allows a loan to an accounting firm to be interest only for its whole term where lending is no more than 2.5 times the firm's earnings before interest, tax, depreciation and amortisation (EBITDA), and for no more than two years where it is higher.⁹ The table in Matching the debt to the asset gives more examples, each a lender's own term with conditions, not a market norm.

Repaying principal is not an expense, so it does not reduce profit. It does use cash: a business can report a healthy profit and still miss its loan payments, because the payments are set by the loan's schedule and the cash by the business. Lenders measure the gap with debt service cover: the cash available for the loan, usually EBITDA less tax paid and often less capital spending and distributions, divided by a year of interest and scheduled principal, with the loan agreement's definition governing. Below 1.00, the business cannot pay the loan from what it produces. Our second model takes cash after tax and the spending the business needs, the stricter measure; the stress tester in Part 5 divides EBITDA itself by the payments, so the same business shows higher cover there. Part 4 sets out the other measures that size a loan.

Banks and the other authorised deposit-taking institutions (ADIs: banks, building societies and credit unions) must look at the same question under stress. Under APRA's credit risk standard, a bank's assessment of a business borrower must consider, where relevant, its capacity to repay under scenarios such as higher interest rates and lower income or cash flows, and must rest "primarily on the strength of a borrower's repayment capacity" rather than on its security (APS 220, paragraphs 42, 45(c) and 46).¹⁰ Covenants, promises in the loan agreement such as keeping a ratio above a set level, add a test that runs for the life of the loan. Under the Banking Code of Practice, which binds the banks that have adopted it, called Code banks here, a standard form small business loan may use financial covenants as a trigger for action on a default in two cases: loans for property development, and specialised lending, a list that includes "tailored cash flow lending" (paragraph 91).⁷ Negotiated and larger facilities are outside this, and covenant defaults are usual there; Part 5 explains how lenders set and test them. The effect is to turn a question about returns into a question about cash: not what the owners will have earned by the end, but whether the business can pay every instalment until then.

Our second model shows the difference. It buys a hypothetical business for $2 million, five times its yearly EBITDA of $400,000, and assumes 60% of those earnings is left as cash for the lenders and the owners, the rest going on tax, equipment and the other spending the business needs to keep running. The debt costs 8% and is repaid in equal instalments of principal over seven years. In the base case, earnings grow 3% a year and the business is sold after five years at the same five times earnings. In the worst case, earnings fall 30% in the first year and stay there, and the business sells at four times earnings.

Illustrative model

Model: leverage, higher returns and higher losses

The four scenarios borrow different shares of the same $2 million price. Compare the base case with the worst case, then change the term of the debt, the rate or the size of the fall.

No debtDebt 30%Debt 50%Debt 70%

Without JavaScript the model shows one scenario: Debt 50%.

The business

Purchase price

$2,000,000

−+

EBITDA a year

$400,000

−+

Share of earnings left as cash

60%

−+

Years until it is sold

5 years

−+

The debt

Debt, share of the price

50%

−+

Interest rate

8.00%

−+

Years to repay the debt

7 years

−+

The two cases

Base case: earnings growth a year

+3.0%

−+

Worst case: fall in earnings, then no growth

−30%

−+

Worst case: fall in the sale multiple of earnings

1.00

−+

Base case 18.3%worst case −6.8%

Base case: return on the owners' money, a year

18.3%

Worst case: return on the owners' money, a year

−6.8%

Worst case: money back, times what went in

0.72×

Worst case: lowest debt service cover

0.75×

Borrowing 50% of the price moves the owners' return from 14.5% to 18.3% a year in the base case, and from −0.5% to −6.8% in the worst case.

In the worst case, cash flow falls short of the loan payments, and the owners must put in $160,000 over 5 years to keep the loan current.

In the worst case, the sale returns $834,000 to the owners after the debt is repaid, against $1,965,000 in the base case.

Leverage: higher returns, higher losses

Base case

18.3%a year

Return on the owners' money

Money back, times what went in

2.24×

Lowest debt service cover

1.08×

To the owners at sale

$1,965,000

Extra money put in

None

Worst case

−6.8%a year

Return on the owners' money

Money back, times what went in

0.72×

Lowest debt service cover

0.75×

To the owners at sale

$834,000

Extra money put in

$160,000

Return on the owners' money, by how much is borrowed

Base caseWorst case

Illustrative model

The same business at four levels of debt, in a base case and a worst case

Line	No debt	Debt 30%	Debt 50%	Debt 70%
The owners' money at the start	$2,000,000	$1,400,000	$1,000,000	$600,000
Debt	$0	$600,000	$1,000,000	$1,400,000
Base case: return a year	14.5%	16.3%	18.3%	21.5%
Worst case: return a year	−0.5%	−3.4%	−6.8%	−12.8%
Base case: money back, times	1.76×	1.97×	2.24×	2.50×
Worst case: money back, times	0.98×	0.85×	0.72×	0.62×
Base case: lowest debt service cover	n/a	1.79×	1.08×	0.77×
Worst case: lowest debt service cover	n/a	1.26×	0.75×	0.54×
Worst case: extra money the owners put in	$0	$0	$160,000	$560,000
Base case: to the owners at sale	$2,251,000	$2,080,000	$1,965,000	$1,851,000
Worst case: to the owners at sale	$1,120,000	$949,000	$834,000	$720,000
Worst case: lender's shortfall	$0	$0	$0	$0

Returns are a year, over the five years to the sale. Money back counts every dollar the owners put in, including any shortfall they make good, against every dollar that comes back to them. Debt service cover is the cash available divided by a year of interest and principal, at its lowest point; with no debt there is nothing to cover. Amounts are rounded to the nearest $1,000.

The two tables below keep the rest of the worst case and change how much is borrowed and how far earnings fall.

Illustrative model

Worst case: return on the owners' money a year, by the share of the price borrowed and the fall in earnings

	Fall in earnings
Debt, share of the price	0%	−15%	−30%	−45%
0%	8.6%	4.3%	−0.5%	−6.0%
30%	8.8%	3.1%	−3.4%	−11.4%
50%	9.0%	1.8%	−6.8%	−17.9%
70%	9.4%	−0.5%	−12.8%	−30.1%

In every column earnings stop growing and the business sells at four times earnings rather than five. Losses carry a minus sign and are shown in red.

Illustrative model

Worst case: extra money the owners must put in over five years to keep the loan current

	Fall in earnings
Debt, share of the price	0%	−15%	−30%	−45%
0%	$0	$0	$0	$0
30%	$0	$0	$0	$2,000
50%	$0	$26,000	$160,000	$340,000
70%	$200,000	$380,000	$560,000	$740,000

The shortfall each year between the cash the business produces and the loan's interest and principal, added up over the five years to the sale and rounded to the nearest $1,000.

Where the debt sits: arranging borrowing in an investment company

An investment company with more than one holding decides not only how much to borrow but where the debt sits: at the holding company, in each business, or against a property. The choice sets what each lender can reach if something goes wrong, and so how far one failure can travel. A lender to one company is owed money by that company, and can reach another company's assets only if it has been given a right to them: a guarantee from that company, or security over its assets or shares. A cross-default clause lets a lender treat a default elsewhere in the group as a default on its own loan; a negative pledge stops a company giving security to anyone else without the lender's consent; and all moneys wording makes security given for one loan secure everything that company ever owes the lender. All of these are contractual. The law can reach past a wall between companies in a liquidation, as Structure B explains.

The five structures below, and the holding company borrowing that lenders often propose instead of the second, are illustrations for a hypothetical investment company in its early years, with two or three holdings. They describe choices, not the arrangements of any firm, ours included.

Structure A: all equity

All equity is the simplest structure and the most expensive. The shareholders fund the holding company, which owns each company outright, and no lender appears anywhere. No lender can take anything and no covenant can be breached, but the owners provide the whole price and carry every loss, with no lender's assessment testing the plan before they commit, and the first model shows what it costs them in return. All equity suits, in our view, an investment company's first holdings, while they are small relative to its capital and their cash flows untested, and any holding whose cash flow is too uncertain to carry repayments.

Structure B: debt in each company, with no cross guarantees

Illustrative model

Structure B: each company borrows for itself, with no cross guarantees

Arrow: from one party to the other, as its label says

A diagram of six parties: Shareholders; Holding company, No borrowing; Company A; Company B. To one side: Lender to A; Lender to B. How they connect:

How they connect

1. 1Shareholders to Holding company Equity
2. 2Holding company to Company A Owns 100%
3. 3Holding company to Company B Owns 100%
4. 4Lender to A to Company A Loan, secured on A only
5. 5Lender to B to Company B Loan, secured on B only

No guarantee runs between the companies or from the holding company, so neither lender has a contractual claim beyond the company it lent to.

Each company borrows against its own assets and earnings, and no company guarantees another's loans. If Company A defaults, its lender can enforce against A but has no contractual claim on Company B or on the holding company. That wall is the point of the structure, and it has limits of two kinds.

The first is that lenders ask for the wall to come down. A guarantee from the holding company, a director's personal guarantee, or a clause in B's loan that treats A's default as its own would each carry A's problem across, and a bank lending to one company in a group will usually ask for exactly those: a guarantee from the holding company and the directors, often from the other companies too, and it will size and price the loan on the group if it gets them. Structure B is therefore available only where each company can carry its own loan on its own figures, and the price of refusing the guarantees is a smaller loan, a higher margin or a different lender. The wall is strongest with different lenders: one bank lending to both companies will usually document them together, with cross-default, all moneys security and a right of set-off between accounts, unless those are negotiated out. And whatever the holding company has lent to A ranks as an unsecured creditor, usually behind A's lender under a subordination deed, so its exposure to A is its equity plus its loans.

The second is the law. A liquidation can reach past the contractual position. A holding company can be made liable for debts a subsidiary incurred while insolvent, where there were reasonable grounds to suspect it and the holding company or one of its directors was aware of them, or a holding company in its position would be expected to be (Corporations Act 2001, section 588V);¹¹ with the same people on both boards, that is usually the case. A court can also, in some cases, pool the assets of related companies in liquidation (section 579E).¹² And directors must not let a company trade while insolvent, a duty ASIC notes applies even to a sole director and shareholder (section 588G).¹³ The wall stops a lender; it does not stop a liquidator. It is a reason to act early, not a reason to wait.

Debt at the holding company

The other way to borrow for a group is at the holding company itself, which is what a lender to an investment company most often proposes. A lender there is repaid from the dividends and loan repayments the subsidiaries send up, after each subsidiary's own lenders, suppliers, employees and the tax office have been paid, so on its own it sits behind everyone else: structural subordination, lawyers call it. It usually asks for what cures that: guarantees and security from each subsidiary, a mortgage over their shares, and covenants measured across the group. The facility can be larger and cheaper than the sum of B's loans, because it is backed by everything. The price is that there are no walls. A failure anywhere is the holding company lender's business everywhere.

Structure C: the property held apart

Illustrative model

Structure C: the property held apart and leased to the business

Arrow: from one party to the other, as its label says

A diagram of six parties: Shareholders; Holding company; Operating company; Property company. To one side: Property lender; Business lender. How they connect:

How they connect

1. 1Shareholders to Holding company Equity
2. 2Holding company to Operating company Owns 100%
3. 3Holding company to Property company Owns 100%
4. 4Operating company to Property company Rent, under a lease
5. 5Property lender to Property company Loan, registered mortgage
6. 6Business lender to Operating company Overdraft, general security

The premises sit in a company of their own and are leased to the operating company. The property lender lends against the building and the rent, the business lender against the business, and each sees a clearer risk. A building can usually carry longer debt than a business, because its value does not depend on one team's skill. The weakness is the lease that ties the two together: if the business fails, the property company loses its tenant and the rent its loan was sized on, and lenders price that link. Their lease doc products, which lend on the rent rather than the borrower's full accounts, show how far: the Commonwealth Bank's (CommBank's) Lease Doc loan requires an arm's length lease between independent parties (page undated, read 12 October 2026),¹ and Bendigo Bank's lease doc brochure, coded 03/26, a lease to a third party, not a related entity.¹⁴ Part 4 sets out both products' terms. A property let to the owners' own business is therefore usually assessed on that business's accounts, like any other loan to it, and often guaranteed by it. Separate ownership is not separate risk when the landlord and the tenant have the same owners.

Structure D: vendor finance in an acquisition

Illustrative model

Structure D: vendor finance in an acquisition, ranking behind the bank

Arrow: from one party to the other, as its label saysDashed line: a promise, not money

A diagram of four parties: Buyers; Buyer company, Buys the business. To one side: Bank; Vendor. How they connect:

How they connect

1. 1Buyers to Buyer company Equity, $800,000
2. 2Bank to Buyer company Senior loan, $900,000
3. 3Vendor to Buyer company Vendor loan, $300,000
4. 4Buyer company to Vendor Price, less the vendor loan
5. 5Vendor and Bank (dashed line) Ranks behind the bank

The vendor lends part of the price, $300,000 in the first model's bank and vendor scenario, to be repaid from the business over the years after completion, and agrees to rank behind the bank. The model repays it from the sale after five years, longer than most vendor loans run: in practice one is often shorter than the bank's loan, sits beside a deferred or earn-out payment tied to the handover, and the buyer asks for the right to set what it is owed under the warranties against it. For the buyers, vendor finance reduces the equity needed and keeps the seller interested in a good handover. For the vendor, it is a loan to a business it no longer controls, behind a bank that is repaid first: in the model's scenario it is repaid in full unless the value at the sale falls by more than 40%, the point at which the owners' equity is gone, and with a mezzanine lender ahead of it as well, as the opening showed, it can lose everything. Two terms matter most, in our view: what the vendor can do if a payment is missed, which its deed with the bank will usually limit, and whether the buyers have guaranteed it personally.

The illustration is a purchase of the business's assets. Where the company's shares are bought instead, the bank will want security over the company's own assets for a loan that paid for its shares, which the law treats as the company giving financial assistance to buy itself. That is allowed only if it does not materially prejudice the company, its shareholders or its ability to pay its creditors, or its shareholders approve it under the Act's procedure (Corporations Act 2001, sections 260A and 260B).¹⁵ It is routine, and the approval takes time, so it belongs in the timetable.

Structure E: private credit for growth

Illustrative model

Structure E: a private credit loan for growth, ranking behind the bank

Arrow: from one party to the other, as its label saysDashed line: a promise, not money

A diagram of four parties: Shareholders; Operating company, Expanding. To one side: Bank; Private credit fund. How they connect:

How they connect

1. 1Shareholders to Operating company Equity
2. 2Bank to Operating company Senior loan, first ranking
3. 3Private credit fund to Operating company Mezzanine loan, second ranking
4. 4Private credit fund and Bank (dashed line) Intercreditor deed
5. 5Operating company and Private credit fund (dashed line) Tighter covenants

A private credit fund lends behind the bank to pay for growth the bank will not fund on its own, such as a second site, a new production line or an acquisition. This is one form private credit takes, shown here because it changes the order of payment. In our reading of the market, a business that meets a private credit term sheet in Australia is more likely to meet a senior loan, sometimes the whole facility in place of a bank, at a higher rate and on tighter terms; the Reserve Bank notes that Australian private credit funds "are more exposed to real estate, including construction and development".¹⁶ Private credit is a small part of Australian business lending, but it is growing (Part 3 has the figures).¹⁷

The two lenders sign an intercreditor deed, which sets who is paid first and what each may do on a default. It usually stops the junior lender enforcing for a set period after a default, lets the bank block interest to it while the bank is unpaid, requires it to hand over anything it receives meanwhile, and lets the bank sell the business free of the junior lender's security. APRA's credit risk standard binds ADIs (APS 220, paragraph 2),¹⁰ so a fund that is not one sets its own standards, and the price of its flexibility is more than the rate: tighter covenants, more frequent reporting and often a share of the upside. In the first model's stretched scenario, the mezzanine layer at 14%, almost twice the bank's 7.5%, helps lift the interest bill to $136,000 a year. Growth paid for this way needs the growth to arrive on time.

How the structures compare

Each structure also has tax and duty consequences that this article does not cover and that usually decide the matter as much as risk: whether, and to which company, the interest is deductible; transfer duty when a property changes hands, and landholder duty when a company that holds land does (in New South Wales, under the Duties Act 1997); the treatment of loans between a private company and its shareholders or their other companies (Income Tax Assessment Act 1936, Division 7A); and the capital gains tax position if the structure is changed later. Those are questions for an accountant and a lawyer before a structure is chosen, because a structure that is right on risk can be wrong on tax, and changing it later can cost duty a second time.

Threefold Capital interpretation

The structures compared

Structure	Equity needed	How far a failure can reach	What can undo the protection
A. All equity	All of the price	Each loss stays in its company, but every loss falls on the owners	Nothing: there is no lender
B. Debt in each company	Less, company by company	The company that borrowed, and its lender	A guarantee from the holding company or a director, a cross-default clause, or a liquidator's claim against the holding company for the subsidiary's insolvent trading (section 588V)
Debt at the holding company	Least, as one facility is sized on the group	Everywhere: the lender is repaid from every company and usually holds guarantees and security from each	Nothing to undo: the protection was traded for the facility
C. Property held apart	Less, as the property carries its own loan	Through the lease: if the business fails, the property loses its rent	A guarantee of the property loan by the operating company or a director
D. Vendor finance	Less, by the vendor's share	The vendor's loan first, once the owners' equity is gone	A guarantee of the vendor loan by the buyers
E. Private credit	Less, at a higher cost	Both lenders, the private lender first	Covenants tight enough to give the lenders control early

A minority investment: what the investor can and cannot control

An investor that owns a minority of a company does not decide how much it borrows. The board does, answering to the majority. In the illustration below, an investment company holds 25% of a business whose founders hold the rest, and the business borrows from a bank against its own assets. The percentages are arbitrary; the illustration describes no holding of ours. The investor's equity ranks behind the bank's loan just as the owners' does in our first model, so if the business borrows heavily and fails, the investor's money is lost before the bank's, over a loan it had no say in unless it negotiated one.

Illustrative model

A minority investment: what the investor holds, and what it can ask for

Arrow: from one party to the other, as its label saysDashed line: a promise, not money

A diagram of four parties: Investment company; Founders; Investee company. To one side: Investee's lender. How they connect:

How they connect

1. 1Investment company to Investee company Equity, 25%
2. 2Founders to Investee company Equity, 75%
3. 3Investee's lender to Investee company Loan, with security
4. 4Investee company and Investment company (dashed line) Information and consent rights

What a minority investor can negotiate, usually in a shareholders' agreement or the company's constitution, is information and a veto over the decisions that change the company's risk. Four rights matter most where borrowing is concerned.

• Information. Regular accounts, the annual budget, the loan agreements and the covenant reports the lender receives, and prompt notice of any breach. Without a contract the position is thin but not empty: shareholders with at least 5% of the votes in a small proprietary company can direct it to prepare a financial report and directors' report for a year and send them to all shareholders, audited if the direction says so (Corporations Act 2001, section 293),¹⁸ and ASIC can direct the same; otherwise, as ASIC notes, only some small proprietary companies must lodge financial reports.¹⁹
• Consent over new debt. A limit above which new borrowing, new security, guarantees by the company and changes to existing loans need the investor's approval. We think a limit tied to earnings, such as debt as a multiple of EBITDA, works better than a dollar figure, which stays fixed while the business changes.
• A seat on the board, or an observer. A seat gives information and a voice, not control, and a director nominated by an investor owes duties to the company, not to the investor: directors must act in good faith in the best interests of the company (Corporations Act 2001, section 181).²⁰
• A say in new equity. A right to take up new shares before they are offered to anyone else, so that a capital raising to repair the balance sheet does not dilute the investor without an offer.

Two terms sit outside the shareholders' agreement. A lender may ask every large shareholder to guarantee the loan or to mortgage its shares, so the first thing to settle is whether the investor will guarantee debt it does not control; in our view it should not. And loan agreements commonly stop dividends while a covenant is breached or cover is below a set level, so the lender, not the board, can switch the investor's income off. A loan from the investor to the company will usually be subordinated to the bank's as well.

What a minority investor cannot control matters as much. It cannot change the lender's terms or decisions, stop the lender enforcing on a default, or make the majority put in more money when the business needs it. A consent right is a veto, not a plan: it can stop a bad loan but cannot make a good one happen, and set too wide it can hold up a business that needs to move quickly. The line, we think, belongs where a decision changes the company's risk, not where it changes the day-to-day running of the business.

Matching the debt to the asset

The safest debt is repaid by the asset it paid for, within that asset's life, from what the asset earns. The danger is a mismatch in either direction: debt that outlives what it paid for, or debt that falls due before the asset has earned the money.

Threefold Capital interpretation

Matching the debt to what it pays for

What the money pays for	Debt that matches it	Published terms, for example	The mismatch to avoid
Stock and money owed by customers	A facility that rises and falls with trade, reviewed each year	Judo Bank's line of credit has a "12-month term with annual review" (July 2026). Macquarie's business overdraft is repayable on demand (May 2026)	Paying for long-lived assets with a facility that can be cut at its review
Equipment and vehicles	A loan repaid over the asset's working life and secured on it	NAB says chattel mortgage repayments can be set over "a range of terms, usually two to five years" (page undated, read 12 October 2026)	A large final payment that falls due when the equipment is worth little
A business, mostly goodwill	A loan repaid from cash flow over a few years, with equity or vendor finance for the rest	Judo Bank's acquisition loan: up to 100% of acquisition costs, to $15 million, terms up to 15 years, interest only for up to five (June 2026). Macquarie: lending against a rent roll, up to 70% of its value, must amortise to 60% within five years (May 2026)	Interest-only debt whose repayment depends on selling or refinancing the business
Property let to tenants	A longer loan sized on the rent and the lease	ANZ: terms of up to five years "matched to remaining lease periods" (November 2025). CommBank's Lease Doc loan: up to five years or lease expiry, whichever is shorter (page undated, read 12 October 2026)	A loan that falls due as a major lease ends
A development	A construction loan repaid from the sales	Lenders' published terms are set out in Part 4	Sales that settle after the loan falls due

Each lender's own published terms or maximums on the date shown, with conditions. They are examples, not market norms, and not recommendations of any lender or product.

Refinancing risk

A loan that is not repaid by its own instalments must be refinanced or repaid when it falls due, on the terms available then, which can be very different from the terms at the start. The cash rate rose four times in 2026, from 3.60% to 4.60%,⁵ and Part 3 sets out what that has meant for borrowers who need to refinance. For a small business with a loan from a Code bank, if the business is not in default and the loan is not fully repaid by regular instalments, the bank must give at least three months' notice if it will not extend the loan, and if it does extend or refinance, it is "not required to do so on the same terms" (paragraphs 93 and 94).⁷ Three months is not long to find a new lender in a difficult year, and the Code's small business protections apply only to a business, with its group, inside the Code's limits on turnover, staff and total debt (Part 4 sets them out),⁷ so a group that borrows more can move outside them.

The longer term that eased the 70% structure in our second model has its own cost here. Repaid over 15 years, the loan leaves $933,000 outstanding at the sale, against $400,000 when it is repaid over seven: easier to carry and larger to clear. We believe a refinancing should be planned as if it will fall in a bad year, with a clear idea of which lender would be asked and what it would want to see.

Covenant headroom

Headroom is the distance between where a covenant ratio stands and where the covenant is set. Our second model measures debt service cover on the cash left after the business's own needs, and with 50% borrowed that cover is 1.08× at its lowest point in the base case. Against an illustrative minimum of 1.25 times applied to this cash measure (the model in Part 5 uses the same level, but divides EBITDA before tax and capital spending by the payments, a more generous test), the structure would be in breach in its first year with nothing gone wrong. Headroom costs money, whether as less debt, a longer term or a higher rate for looser terms. We believe it is worth paying for, because, depending on the loan agreement, a breach can give the lender rights even when every payment has been made, and those rights arrive when the business can least afford them.

Security and guarantees

Security decides what a lender can take, and a guarantee decides who else it can pursue; Part 6 sets out the law. For the structure, two points matter. A general security agreement over all of a company's present and after-acquired property, registered on the Personal Property Securities Register, reaches all of the company's personal property,²⁵ which does not include land.²⁶ Between lenders with perfected security over the same assets, the default rule ranks the earliest registration first (Personal Property Securities Act 2009, section 55, as the register summarises it),²⁷ so a junior lender's place in the stack depends on registration as well as on its deed with the senior lender, which can set an order the register does not.

A director's personal guarantee reaches past the company to the director's own assets, which is why lenders ask for one even when they hold security: CommBank lists a director guarantee among the security for its Lease Doc loan,¹ and ANZ says that for its unsecured business loans "a personal guarantee may still be required from all company directors" (page updated 8 October 2026).² A Code bank will not accept a guarantee until the third day after giving the guarantor the information the Code requires, unless an exception applies (paragraphs 112 and 113).⁷ We think those days are best used to price the guarantee: its limit, how long it lasts, and what it would cost if called.

Raising equity within the law

Equity is the layer every other layer relies on, and raising it is regulated. Chapter 6D of the Corporations Act 2001 governs fundraising through the issue or sale of securities. A proprietary company, one with no more than 50 shareholders who are not employees, can raise money from its existing shareholders and employees, and from the general public only if the fundraising does not require a disclosure document, such as a prospectus.³ Two exemptions matter most for a private company.²⁸

• Small scale personal offers. No disclosure document is needed for personal offers of a company's securities if they do not result in securities being issued to more than 20 persons, or in more than $2 million being raised, in any 12 month period (Corporations Act 2001, sections 708(1) to 708(7)).⁴ A personal offer is one that only the person it is made to can accept, made to someone likely to be interested in it because of previous contact, a professional or other connection, or things they have said or done (section 708(2)); issues under another exemption, under a disclosure document, or to people outside Australia are not counted (section 708(5)). An offer made under this exemption must not be advertised.²⁸
• Offers to particular investors. Nor is one needed where the offers are made to people "presumed not to need disclosure because of their financial capacity, experience, association with the issuer or wholesale status" (sections 708(8) to 708(12)).²⁸ Each category has its own tests in those sections, which this article does not attempt to summarise.

These are general rules, stated from the Act and ASIC's guidance; whether an exemption applies to a particular offer turns on its facts and is a question for a lawyer. This article is not an offer of shares in Threefold Capital or in any other company, or an invitation to apply for them.

The rules carry a lesson beyond compliance. Equity takes the first loss, so an investor asked to provide it has reason to expect at least the information a lender would receive, including the worst case and the money the owners might have to add. A structure that cannot be explained plainly to a prospective shareholder is unlikely to be explained well to a lender.

The Threefold Capital perspective

The principles below are our views on the use of debt. They describe how we think, not any transaction or policy of the firm, and they are not recommendations to anyone else. Each rests on the models above or on the earlier parts of this series.

1. Debt should be serviceable from the cash an asset produces in a bad year, which in our models means after a fall of about a third. That fall is what leaves the owners of the first model's stretched structure nothing after interest. In the second model, borrowing 70% of the price gives the best base case return and cannot be repaid from the business's cash even there, while 30% is still covered 1.26× when earnings fall by 30%. Banks must consider a borrower's capacity under higher rates and lower cash flows;¹⁰ owners have reason to test harder, because they lose first.
2. A lender's maximum is not a target. Published maximums, such as up to 100% of acquisition costs²¹ or several times EBITDA,⁹ describe what a lender will consider under its conditions, not what a particular business can repay. In our model, debt of 3.5 times earnings fails on cash flow in its base case.
3. Headroom is worth paying for. A covenant is set against the plan, and a business lives in the range around it: with 50% borrowed in our second model, cover is below an illustrative 1.25 times before anything has gone wrong. Less debt, a longer term or a slightly higher rate for looser terms are each cheaper than a breach, which hands options to the lender at the worst moment.
4. The structure should let one asset fail without taking the others. Separate companies, separate lenders and no cross guarantees, as in Structure B, keep a failure where it starts, as far as the lenders and the law allow. In our view that wall is worth a higher margin, and it should be defended each time a loan is renewed.
5. A personal guarantee is equity at risk, and should be counted as money put in. A guarantee turns a company's debt into a director's own, and lenders' pages show how routinely one is asked for.^1,2 In our first model's 30% fall, a guarantee of the vendor loan would have turned the vendor's loss of $300,000 into the owners'. Counted as money put in, a guarantee changes both the equity in a structure and the return on it, and anyone asked to give one should know its limit, its term and its cost if called before comparing interest rates.
6. Debt should be repaid by the asset it paid for, within that asset's life. Lenders' own terms already match short facilities to working capital and lease terms to property. The hard case is goodwill, which nobody will buy apart from the business, so debt against it should be repaid from cash flow rather than left for a sale, which is where the 15-year term in our second model leaves most of it.
7. Equity should be patient enough to outlast a refinancing cycle. Loans that do not amortise must be refinanced on the terms of the day, and a Code bank need give a small business that is not in default only three months' notice that it will not extend a loan not fully repaid by instalments.⁷ We believe equity in a borrowed structure should be committed for longer than its debt, and should not depend on a sale in a particular year.
8. The worst case belongs beside the base case, with the owners' top-ups beside it. In our second model, borrowing 70% rather than 30% lifts the base case return from 16.3% to 21.5%, lowers the worst case from −3.4% to −12.8%, and leaves $560,000 to find along the way. The worst case belongs first in any proposal to borrow, and the money the owners might have to add beside it.

Conclusion

This series began with a single loan and what it costs. Part 1 traced a loan's rate back through the cash rate, the lender's funding and the capital it must hold. Part 2 applied that to the loan most Australians know, the home loan. Part 3 looked at what four rate rises in 2026 have meant for borrowers and lenders. Part 4, Part 5 and Part 6 followed a business loan from the measures that size it, through the decision to make it, to the security behind it and the law that applies when it goes wrong.

This part has taken up the question the others lead to: when, and how, a business or an investor should borrow at all. The answer is the same from both sides of the table. A bank must assess a loan primarily on the borrower's capacity to repay it, without leaning on its security,¹⁰ and the owners have more reason than the bank to do the same, because they lose first. Debt adds value when the asset earns more than the debt costs, with room to spare; when its repayments fit the cash the asset produces in a weak year; when its term fits the asset's life; and when the structure keeps one failure contained. It destroys value when any of those fails, and that tends to happen when the borrower can least afford it.

In our view, using debt well is mostly a matter of declining the debt that is offered but not needed. The models in this series are there to make that judgement concrete: change an assumption, see who pays, and decide before a lender has to.

Method and sources

Cut-off. This article uses data, rules and lenders' published terms available on or before 12 October 2026. Lenders' pages were read on that date, and the sections of the Corporations Act 2001 on 13 October 2026.

Sources. Official figures come from the Reserve Bank, prudential rules from APRA's standard APS 220, and the commitments of banks to small businesses and guarantors from the Banking Code of Practice. The Corporations Act points come from the sections named in the text, read in a consolidated text, and from ASIC's guidance where the text says so; the priority rules from the Personal Property Securities Register's summary of the Personal Property Securities Act 2009. Each figure is numbered to its source below.

Lenders' terms. Every lender figure is that lender's own published term or maximum, on its own page, dated as the page is or as we read it, with the lender's conditions; none is a market norm or a recommendation.

Market practice. What lenders usually ask of a group, what an intercreditor deed usually contains, the usual shape of a vendor loan, distribution lock-ups and limits on how often owners may make good a shortfall are described from practice in the Australian market, not from a published source, and any particular agreement can differ.

Our calculations. Where we derive a figure from the models, the text says so, and the arithmetic is here. The $22,500 is the 2.5 percentage point gap between the business's 10% return and the bank's 7.5% rate, on $900,000 of debt; spread over the owners' $1,100,000, it adds 2.05 percentage points, which we call about two. The bank's 7.5% is 2.9 percentage points above the 4.60% cash rate. In the first table, at $100,000 of cash flow and 70% borrowed, interest of $105,000 exceeds the cash flow by $5,000 a year, which over five years is the $25,000 the owners put in. Debt of $1,400,000 is 3.5 times EBITDA of $400,000. In the stretched structure, interest of $136,000 is 68% of the $200,000 cash flow, so a fall of 32%, about a third, would leave nothing for the owners. In the second model, borrowing 70% rather than nothing raises the base case return by 7.0 percentage points, from 14.5% to 21.5%, and lowers the worst case by 12.3, from −0.5% to −12.8%. The lowest debt service cover in the base case falls in the first year, because earnings grow and the loan shrinks after it.

The models. Both businesses are hypothetical, and every setting is our assumption, not a quote, a forecast or an industry average. Each model is written twice, for the page and for the controls, and the two are tested against each other on every scenario and thousands of other settings.

Illustrative model

What the two models assume

Model	Setting	Assumption
Capital structure	The purchase	$2 million, sold after five years, for what it cost except in the last scenario, where the value is 30% lower
Capital structure	Cash flow	$200,000 a year before interest, the same every year; where it does not cover the interest, the owners pay the difference
Capital structure	The debt	Bank 7.5%, mezzanine 14% and vendor 6% a year, all interest only and repaid from the sale in that order
Capital structure	Left out	Tax, fees, covenants and any repayment before the sale
Leverage	The business	Bought for $2 million, five times EBITDA of $400,000, with 60% of EBITDA left as cash for the lenders and the owners, and sold after five years
Leverage	The debt	8% a year, repaid in equal instalments of principal over seven years, with any balance repaid from the sale
Leverage	Base case	Earnings grow 3% a year, and the business sells at five times earnings
Leverage	Worst case	Earnings 30% lower from the first year, with no growth, and the business sells at four times earnings
Leverage	Left out	Tax on the owners, fees and covenants. The share of earnings left as cash is held at 60% whatever the debt, so the model does not credit the borrower with the tax saved on interest, which leans against debt

Every setting can be changed in the models above.

Labels. Each table, diagram and model carries one of four labels. Reported data: figures published by the source named. Threefold Capital interpretation: our reading of the evidence. Illustrative model: hypothetical figures from our models or illustrations. Forward-looking scenario: a planning case, not a forecast.

Limitations. The models leave out tax, including the deduction a borrower generally gets for interest, which makes debt cheaper after tax than the rates shown and the second model's shortfalls smaller. They also leave out fees, intercreditor terms and the way lenders, owners and customers respond to one another in a downturn, and real structures rarely look as tidy. Where the text relies on a regulator's summary rather than the Act, it says so, and the Act's own wording governs.

Corrections. If you see an error, please tell us at threefoldcapitalau@gmail.com. We correct published articles and record what changed at the top of the page.