The Missing Rung

Almost half of Australia’s 2.5 million small businesses are not breaking even, and the sector has contracted for three consecutive years. Meanwhile up to 1.3 million workers may need to change what they do by 2030. Those people have always gone into small business, and Australia has made it unusually hard to get there. This memo sets out what stops them, what other countries built instead, and what a government would do about it.

1. The numbers

• 2.5 million small businesses in Australia; 2.8 million actively trading businesses at 30 June 2025.
• Almost half of those small businesses are not breaking even.
• One in four owners is funding operations from personal savings.
• 95 per cent of companies entering insolvency have fewer than 20 employees.
• Survival for businesses turning over under $50,000 fell from 78 per cent in 2018 to 48 per cent in 2021.
• Solo businesses have a 43.3 per cent three-year survival rate.
• Small businesses have declined in net terms for three consecutive financial years, and in 2024–25 every category under 200 employees fell.
• Fewer than 5 per cent of small business loans are unsecured. About half are secured against residential property.
• Lending to small business grew by about $58 billion over the period housing lending grew roughly ten times that.
• Up to 1.3 million Australian workers may need to move into new roles by 2030 as AI reshapes the labour market.
• 65 per cent of the skills required for existing jobs are expected to have changed by 2030.
• 80 per cent of Australian small businesses are using or planning to adopt AI.

2. What is happening to work

Australia is not currently experiencing AI-driven job destruction, and this memo will not claim that it is.

The Department of Employment and Workplace Relations examined every occupation exposed to AI using data to February 2026 and found no significant impact on overall employment growth and no major shift in the types of jobs across occupations. Employment for people aged 20 to 24 grew slightly faster than for those over 25. Employment in software development is up 25 per cent since generative AI became widely available. Jobs and Skills Australia puts only 4 per cent of Australian jobs at high exposure and concludes that generative AI has a greater capacity to augment work than to automate it.

That is the honest position today, and anyone campaigning on imminent mass unemployment is misrepresenting the evidence.

What is coming is a transition

The forward estimates describe something different from the present. McKinsey estimates that up to 1.3 million Australian workers — around 9 per cent of the workforce — may need to transition into new roles by 2030. Pearson finds that 65 per cent of the skills needed for existing jobs will have changed by the same year. Deloitte Access Economics has identified 82 “AI-disrupted” occupations at highest risk of declining employment, and reports that vacancies in some of them have already started to fall.

A transition of 1.3 million people is not a catastrophe. It is roughly the number of Australians who change jobs in an ordinary two-year period. It becomes a catastrophe only if there is nowhere for them to go.

The inversion

Every previous wave of automation hit factories, farms and trades. This one does not.

The occupations most exposed are clerical, administrative and data entry roles, accounting and payroll clerks, and legal support. The occupations least exposed are plumbers, electricians, aged and child care workers, and drivers — work requiring physical presence, manual dexterity and judgement in unpredictable conditions.

Seventy per cent of workers in the least exposed roles are men, and fewer of them hold university qualifications.

That reverses fifty years of advice. The safe path was a degree and a desk. The exposed path was a trade. Both are now the other way around, and no part of the education, training or welfare system has been rebuilt to reflect it.

3. Where displaced people go

When work disappears, people do one of three things: find another job, leave the workforce, or start something.

The first is what policy assumes. The second is what actually happens to a significant share of displaced workers over 50, and it is expensive for the rest of their lives. The third is the one that creates something, and it is the one Australia has made hardest.

The door is jammed

Almost half of Australia’s 2.5 million small businesses are not breaking even. One in four owners is funding the business from personal savings. Ninety-five per cent of insolvencies involve businesses with fewer than 20 employees, and cash flow is the most commonly nominated cause.

Survival rates tell the sharper story. For businesses turning over under $50,000 — which is every business in its first year — survival fell from 78 per cent in 2018 to 48 per cent in 2021. For those between $50,000 and $200,000, from 87 to 64 per cent. Solo businesses now have a 43.3 per cent three-year survival rate.

And the sector has contracted in net terms for three consecutive financial years. In 2024–25 the decline extended to every category of business employing fewer than 200 people.

The businesses that were never started

The businesses that fail are counted. The ones never attempted are not.

The Institute of Public Affairs, analysing ABS business counts, estimates that close to 400,000 businesses were never created between 2005 and 2019 as a result of accumulated regulatory burden, and puts the annual cost of red tape at $176 billion. Those are advocacy estimates from an organisation with a stated position, and this memo treats them as such.

The direction, however, is corroborated by sources with no such position. The OECD finds that business entry and exit rates have declined across most member economies since 2000. The Productivity Commission currently has an open inquiry into regulatory barriers to business dynamism in Australia, reporting late in 2026.

Something is stopping people from starting. The question is what.

4. The fifty-hen cliff

The clearest example of what stops people is small, specific, and typical.

In Australia you may keep hens and sell the eggs. Under 50 layers — about 240 eggs a week — selling direct from the gate, you are exempt from the accreditation and stamping regime in New South Wales, Victoria, South Australia, Western Australia and Tasmania. Tasmania’s threshold is tighter again: under 20 dozen a week, and only to friends, family and colleagues.

Cross that line and you carry FSANZ Standard 4.2.5 in full. Accreditation. A unique producer stamp on every individual egg. Traceability. Auditing. Salmonella control. The same standard as a farm running 500,000 birds.

There is nothing in between. A person with 200 hens is regulated as an industrial producer, on rules written for industrial producers, at costs set by what an industrial producer can absorb. Councils add their own layer on top — flock limits, setbacks, permits, noise — varying by local government area.

The author of this memo attempted to start an egg business and abandoned it at exactly this point.

The ladder is being pulled up

Egg Farmers of Australia states on its own policy page that, following the FSANZ review, it will be “requesting that a recommendation be made that state exemptions not be permitted and that all eggs be stamped”.

That is the established industry asking government to remove the only rung a new entrant can stand on. It requires no interpretation and no characterisation. It is published.

Victoria has already moved in that direction, with mandatory poultry welfare regulations from the first half of 2026 applying to all poultry owners including backyard keepers, converting guidance into law.

The same shape, everywhere

This is not a story about eggs. It is the shape of Australian regulation, and it appears wherever anyone tries to start small.

Between a mortgage and a rental queue, there is nothing. Between sole trader and real employer, there is nothing — which is why owners avoid the one-to-four employee band entirely, staying solo with contractors or jumping straight to five and above. Between a backyard and an industrial operation, there is nothing.

Australia has removed the middle rung from almost everything. The rule is drafted at industrial scale and applied at kitchen-table scale, and the cost of the first step is set by the standard applied to the largest operator in the industry.

The fix is not deregulation, and this memo does not propose it. The fix is proportion. The rule should scale with the risk, and the risk scales with the size of the operation.

5. What other countries do

Australia is not the only country to have noticed this. Several have already built the rung.

The cottage food laws

All fifty United States now have a cottage food law. Arizona was the last, in 2022.

These laws permit low-risk shelf-stable food — bread, cakes, jams, jellies, honey, granola, roasted nuts, spice blends, dry mixes — to be produced in an ordinary home kitchen, with no commercial fit-out and no inspection, and sold directly at markets, roadside stands and increasingly online.

The sector generated approximately $2 billion in United States revenue in 2024, from an almost negligible base a decade earlier, and the number of registered producers has roughly doubled since 2020.

It is regulated, and that is the point. There is a defined product list drawn on food-science grounds — nothing requiring refrigeration, nothing pressure-canned, a pH below 4.6 for pickled goods to inhibit Clostridium botulinum. There is mandatory labelling: ingredients, allergens, net weight, and a “home produced” disclosure. And there is an annual revenue cap, between $5,000 and $75,000 depending on the state, with a graduation point to a commercial kitchen above it. Around 35 states permit pickled products; roughly 20 permit acidified foods. A handful — Wyoming and North Dakota among them — go further with “food freedom” laws extending to meat, dairy and prepared meals.

Australia has no equivalent. Selling home-baked bread here requires a council-approved kitchen, registration and a food safety supervisor — the requirements written for a restaurant.

Forty-nine states have run this for a decade. It cannot seriously be called a public health risk.

The trading allowance

The United Kingdom introduced a £1,000 tax-free trading allowance in 2017. If gross income from self-employment or casual trading is £1,000 or less in a year, there is no tax, no return, no registration, and no obligation to notify the tax authority at all. Above it, the ordinary self-assessment rules apply. A separate £7,500 allowance operates for letting a furnished room.

In Australia, business income is assessable from the first dollar and an ABN is required to invoice.

The British experience also supplies the argument for widening it. Almost two-thirds of young people there say they want to run their own business. Sixteen per cent have started one. Twenty-eight per cent have a side venture of some kind. The Federation of Small Businesses is now campaigning to double the allowance to £2,000 and rename it the Side Hustle Allowance.

The guarantee

The British Business Bank provides government-backed guarantees and lower-cost funding to lenders. A review found the guarantee enabled lending that 51 per cent of borrowers would not otherwise have received. Germany’s state-owned KfW lends alongside commercial banks to share risk. France’s Bpifrance provides guarantees to the same end. An OECD review of these schemes found they increased small business lending while reducing collateral requirements.

That last finding matters more in Australia than anywhere, for reasons set out in section 7.

6. The frontiers

Removing a barrier lets people into markets that already exist. Opening a frontier creates markets that do not. Four are available to an Australian government, and each is currently either closed, criminal, or unbuilt.

Food, farming and processing

The largest frontier is the one described in section 4, because the barrier is regulatory rather than economic. A cottage food tier, proportionate thresholds for eggs and small livestock, small-scale meat processing, on-farm value-add, and the right to sell what you produce to the people who live near you.

Australian agriculture recorded 10.4 per cent multifactor productivity growth last year, leading every industry in the country, while national productivity fell. The capability is not in question. What is missing is any legal way for a small producer to reach a customer without meeting the standard set for a processor a thousand times their size.

Cannabis

A market of substantial size operates in Australia today, entirely outside the law. Every dollar of it is untaxed, every product unregulated, every dispute settled without a court, and every participant a criminal.

Legalisation does not create that market. It transfers it — to growers, processors, retailers, testers, transporters and accountants who pay tax, meet a standard, and can be sued. It is the single largest one-off transfer of economic activity from the criminal economy to the legal one available to any Australian government, and the party’s position on it is set out separately.

Tobacco

The same logic applies in reverse. Excise drove the legal price of tobacco so high that the criminal supply took the market, and an industry that was legal, taxed and regulated became violent, untaxed and unregulated inside a decade. Firebombings of retail premises are now a routine feature of Australian cities.

Restoring a legal price that beats the criminal one returns the trade to businesses that can be inspected, taxed and held responsible — and returns thousands of retail sites to lawful operators.

Artificial intelligence

The fourth frontier is the one that changes the economics of everything else.

One person with AI tools now performs work that required a team. Bookkeeping, design, copywriting, drafting, customer service, scheduling and analysis — the overheads that made a small operation uneconomic are collapsing. The cost of everything except judgement is falling.

Australians are already moving. Eighty per cent of Australian small businesses are using or planning to adopt AI, with adoption at 69 per cent in property services, 65 in finance and insurance and 61 in business services. AI literacy is now the most in-demand skill in the country, and workers with AI competencies earn a median $143,000 — a 56 per cent premium over the workforce median.

The frontier has two halves. There is AI-enabled business: the sole trader who can now compete with a firm. And there is AI as the business: implementation, integration, training and automation for the other 2.5 million small businesses, almost none of which will ever engage a consultancy.

The Office of AI, and the threshold it needs

On 15 July 2026 the Commonwealth established an Office of AI within the Department of the Prime Minister and Cabinet, to coordinate policy across departments and develop Australian Standards for AI, with legislation expected in early 2027 covering large-scale data centres, AI training and the use of Australian copyright material.

Much of it is sound. The stated purpose includes streamlining approvals and providing regulatory clarity. The data centre obligations — that operators underwrite new power supply and not pass the cost to consumers, pay for their own grid connections and water infrastructure, and minimise water use — put the cost of the infrastructure on the party that profits from it, which is a principle this party applies elsewhere. The position that Australian books, music, art and news should not be used for training without the creator’s consent is defensible.

The risk is not that Australia will stop AI. It is that obligations written for the largest players will be applied down the chain until a five-person firm carries them.

That is the fifty-hen cliff again. A high-risk guardrail drafted for a frontier laboratory will catch a two-person consultancy configuring a chatbot for a plumbing business, unless a threshold is built in deliberately.

Sovereign Australia supports the framework and asks for one addition: a threshold. The obligations that apply to a company training a frontier model should not apply to a sole trader implementing one.

7. Funding without the family home

A person who has decided to start something then has to pay for it. In Australia, that depends almost entirely on whether they own a house.

About 50 per cent of all small business loans are secured against residential property. A further 45 per cent are secured against other collateral. Fewer than 5 per cent are unsecured.

The practical effect is that a person without property cannot borrow to start a business. The most common reasons for rejection are insufficient collateral and a trading history shorter than two years — conditions that describe every new business by definition.

This connects the small business problem directly to the housing problem. When a generation cannot buy a house, it also loses access to the only collateral the banking system accepts for enterprise. The housing crisis is an enterprise crisis with a delay on it.

Banks have made their preference plain. Over the relevant period, lending to small business grew by about $58 billion — roughly one-tenth of the growth in housing lending.

The market has already moved

Non-bank lenders have nearly doubled their share of smaller business lending since 2019, reaching about 27 per cent in 2025, and 47 per cent of small and medium businesses now say they prefer a non-bank lender — a record. Fintech lenders approve in 24 to 72 hours against 21 to 35 days at a major bank, and around 80 per cent of non-bank borrowers are funded within a week.

The demand is demonstrated. What is missing is the mechanism that makes the price reasonable, and three comparable countries already operate it.

What works

The British Business Bank, Germany’s KfW and France’s Bpifrance all place government behind part of the risk rather than lending directly against a house. The British review found the guarantee enabled lending that 51 per cent of borrowers would not otherwise have received, and the OECD found such schemes increase lending while reducing collateral requirements.

Sovereign Australia’s existing commitment — an 80 per cent Commonwealth guarantee with no personal guarantee required — is that mechanism. The international evidence supports it, and the Australian collateral data explains why it is necessary here more than anywhere.

Equity crowd-sourced funding has been available in Australia since 2017, permitting raises of up to $5 million a year through a licensed intermediary. It works, at modest scale, and should be maintained. It is not a substitute for credit.

8. What already works

One part of the existing system does exactly what this memo argues for, and it should be said plainly.

Self-Employment Assistance, which replaced the New Enterprise Incentive Scheme on 1 July 2022, provides twelve months of free mentoring, workshops and accredited training through providers in 51 regions. Participants can complete a Certificate III or IV in Entrepreneurship and New Business. Jobseekers may receive a Self-Employment Allowance for up to 39 weeks and rental assistance for up to 26, and the training counts as an approved activity for 13 weeks — meaning a person on JobSeeker is not required to look for work while building the business. It is open to everyone, not only to jobseekers: about half of participants are existing owners returning for growth mentoring.

Free TAFE is now permanent. The Fee-Free TAFE Skills Agreement delivered $1.5 billion for over 500,000 places between 2023 and 2026, and under the Free TAFE Act 2025 a further $1.6 billion to 2034–35 funds at least 100,000 places a year from 2027, prioritised to people out of work, young people, carers, people with disability and women facing economic insecurity.

These programmes do not hand out grants and do not pick winners. They take a person with an idea and no capital and give them training, a mentor and something to live on while they start. That is the correct model, and it exists.

The gap

Free TAFE’s national priority fields are care, construction, technology and digital, and the VET workforce. Entrepreneurship and small business are not among them.

So a displaced clerical worker is offered free training for a job, and must separately find the much smaller Self-Employment Assistance programme to be trained for starting something. With up to 1.3 million people needing to move roles by 2030, that is the wrong way around.

The expansion

Make small business and entrepreneurship a Free TAFE national priority field, so that training to start something sits alongside training to be employed.

Scale Self-Employment Assistance to the size of the transition. The model is proven and already operates in 51 regions. It is currently sized for an ordinary labour market rather than a structural shift.

Add practical AI capability to both — at the level people actually work at. Not a university course. The bookkeeping of a trades business, the rostering of a café, the quoting of a builder.

This is not a new idea requiring invention. It exists, it works, and it needs to be the size of the problem.

9. Efficiency, not subsidy

The instinct when a sector is struggling is to fund it. Australia has a long record of programmes, grants, offsets and schemes for small business, and a sector that has contracted for three consecutive years.

Grants reach the businesses that can afford someone to write the application. They arrive once. They favour the applicant who is already sophisticated enough to find them, which is rarely the person the policy is written for. And they do nothing about the reason the business could not start.

Removing a barrier is permanent, costs the budget nothing, and reaches everybody. A trading allowance does not require an application. A cottage food tier does not need to be administered by a department. A proportionate egg threshold is a line in a regulation. None of them picks a winner, and none of them expires.

The same distinction runs through the party’s cost-of-living position: subsidise a price and you conceal it; reduce a cost and you change it. Fund a business and you support it for a year. Let it exist and it supports itself for decades.

10. Future-proofing

The question a displaced worker faces is not whether AI is coming. It is what they are legally permitted to try on a Saturday morning with two hundred dollars and a kitchen.

At present, in Australia, the answer is close to nothing. They may not sell the bread. They may not sell the eggs above 50 hens. They cannot borrow without a house. They earn no income before the tax system engages. And the training system will fund them to become an employee somewhere else.

Every one of those is a decision, and every one can be decided differently.

A country that expects 1.3 million people to change what they do this decade, and 65 per cent of skills to change with them, has two options. It can process those people through a welfare system and hope employers absorb them. Or it can make it lawful, cheap and fast for them to build something.

AI is simultaneously displacing people from work and making it cheaper than at any point in history to start something. Whether that is a crisis or an opportunity depends entirely on whether the door is open when they arrive.

11. What Sovereign Australia would do

1. Build the rung. A cottage food tier on the American model — a defined product list, a label, a revenue cap, a graduation point — so a person may lawfully make and sell low-risk food from a home kitchen.
2. Scale the rule to the risk. Proportionate thresholds wherever a standard written for industrial operators catches a small one, beginning with eggs, small livestock and on-farm processing. The rule scales with the risk, and the risk scales with the size of the operation.
3. A trading allowance. Income below a set threshold from a small venture, with no registration, no return and no ABN — on the model the United Kingdom has operated since 2017.
4. Three years to find your feet. No income tax on the profits of a new small business for its first three years, and every dollar reinvested deductible in the year it is spent.
5. One return, not four. Quarterly BAS abolished below the small business threshold and replaced with a single annual return.
6. Cross the employment threshold. The payroll tax cliff removed and a genuinely simplified obligation set for employers under five people, so the step from sole trader to employer is a step rather than a wall.
7. Credit without the family home. An 80 per cent Commonwealth guarantee with no personal guarantee required, on the model of the British Business Bank, KfW and Bpifrance.
8. Get paid on time. Statutory payment terms with interest accruing automatically, so a small business stops functioning as an unsecured, interest-free bank for its largest customers.
9. Open the frontiers. Cannabis legalised and regulated, tobacco returned to the lawful economy through a price that beats the criminal one, and food production opened to the people who produce it.
10. Train people to start, not only to be employed. Entrepreneurship made a Free TAFE national priority field, Self-Employment Assistance scaled to the size of the transition, and practical AI capability added to both.

Nine of these cost the budget little or nothing. The tenth is a training programme that already exists.

12. Summary

Australia has 2.5 million small businesses, almost half of which are not breaking even, and a quarter of their owners are funding operations from personal savings. Ninety-five per cent of insolvencies involve businesses with fewer than 20 employees. Survival for the smallest businesses fell from 78 per cent to 48 in three years, and the sector has contracted in net terms for three consecutive years.

AI is not currently destroying Australian jobs, and the Department of Employment and Workplace Relations has found no significant effect on employment growth. But up to 1.3 million workers may need to move roles by 2030, 65 per cent of required skills will have changed, and the exposure has inverted — the office is at risk and the trades are not.

Those people will need somewhere to go, and small business is where displaced people have always gone. Australia has made it unusually hard to get there. The regulation is drafted at industrial scale and applied at kitchen-table scale: fifty hens and you are an industrial producer, with nothing in between, and the established industry is publicly asking government to remove even that exemption. Fewer than 5 per cent of small business loans are unsecured, so a person without a house cannot borrow. And business income is taxable from the first dollar.

Other countries built the rung. Fifty American states permit home-kitchen food production under a product list, a label and a revenue cap, generating $2 billion a year. The United Kingdom lets a person earn £1,000 from a side venture without telling the tax office. Britain, Germany and France put government behind part of the lending risk instead of demanding the family home, and the British guarantee enabled lending that half of borrowers would not otherwise have received.

Australia already has the best part of the answer. Self-Employment Assistance and Free TAFE take a person with an idea and no capital and give them training, a mentor and something to live on. Entrepreneurship is simply not among Free TAFE’s priority fields, and the programme is sized for an ordinary labour market rather than a structural shift.

None of this requires a grant, a scheme or an agency. It requires the country to decide that a person with an idea, no house and two hundred dollars is allowed to begin.

AI is making it cheaper than ever to start something at exactly the moment more Australians will need to. The only question is whether the door is open when they get there.

13. Sources

1. Australian Bureau of Statistics, Counts of Australian Businesses, including Entries and Exits, and Australian Securities and Investments Commission insolvency statistics — approximately 2.5 million small businesses and 2.8 million actively trading businesses at 30 June 2025; around 14,000 companies entering formal insolvency a year as at January 2026; and 3,556 companies entering external administration in the first quarter of 2025–26 from over 3.6 million registered companies; sections 1 and 3.
2. Council of Small Business Organisations Australia, as reported by Export Finance Australia, World Risk Developments — almost half of Australia’s 2.5 million small businesses not breaking even, and one quarter of business owners using personal savings to continue operating; sections 1 and 3.
3. Australian Securities and Investments Commission insolvency data as reported in SmartCompany (July 2026) — almost 95 per cent of insolvent companies having fewer than 20 employees, and cash flow pressure as the most nominated cause of small business failure; sections 1, 3 and 11.
4. Deloitte Australia, Insolvencies up, back to normal levels, drawing on ASIC insolvency statistics — survival rates for businesses with annual turnover below $50,000 falling from 78 per cent in 2018 to 48 per cent in 2021, and for businesses turning over $50,000 to $200,000 from 87 per cent to 64 per cent; sections 1 and 3.
5. ScaleSuite, Australian Business Statistics 2026, drawing on ABS business counts — a 43.3 per cent three-year survival rate for solo businesses; the compliance layering of Single Touch Payroll Phase 2, superannuation rising from 10 to 12 per cent between 2021 and 2025, Payday Super from 1 July 2026, Modern Award complexity and expanding Fair Work requirements; owners avoiding the one-to-four employee band as a result; and a 0.8 per cent decline in agriculture, forestry and fishing businesses to 170,890; sections 1, 3 and 4.
6. Institute of Public Affairs, Australian small business sector in crisis (April 2026), drawing on ABS Counts of Australian Businesses — net decline in small businesses with one to nineteen employees across three consecutive financial years, all categories under 200 employees declining in 2024–25, 1,260 businesses becoming insolvent in February 2026, an estimated 400,000 businesses never created between 2005 and 2019 as a result of regulatory burden, and an estimated $176 billion annual cost of red tape. The Institute of Public Affairs is an advocacy organisation and these are its estimates; sections 1 and 3.
7. Organisation for Economic Co-operation and Development (2021), and Productivity Commission, Regulatory barriers to business dynamism in Australia (inquiry active as at May 2026, reporting late 2026 or early 2027) — business entry and exit rates declining across many OECD economies from 2000 to 2015, and the current Australian inquiry examining regulatory barriers including the design and operation of corporate and personal insolvency frameworks; section 3.
8. Department of Employment and Workplace Relations, AI and Employment in Australia (using data to February 2026) — no significant impact on overall employment growth and no major shifts in the types of jobs across occupations since generative AI became widely available; employment growth for people aged 20 to 24 slightly exceeding that for people aged 25 and over; software development employment increasing 25 per cent since November 2022; and employment in the occupations most exposed to AI growing 5.6 per cent against 9.5 per cent in the least exposed; section 2.
9. Jobs and Skills Australia, Generative AI Capacity Study, and Reserve Bank of Australia — approximately 4 per cent of Australian jobs highly exposed to AI automation and around 21 per cent facing medium-to-high exposure; the finding that generative AI has a greater capacity to augment work than to automate it; an estimate that around 13 per cent of Australian jobs could be automated by 2050 while more than half could be augmented; and the identification of trades, health care, construction and hospitality as least exposed, with men comprising around 70 per cent of workers in the least exposed roles; section 2.
10. McKinsey & Company, and Pearson, Lost in Translation — up to 1.3 million Australian workers, around 9 per cent of the workforce, potentially needing to transition into new roles by 2030 due to automation and generative AI; and around 65 per cent of the skills needed for existing jobs changing by 2030, with 26 per cent of jobs at high risk in the absence of upskilling; sections 1, 2 and 10.
11. Deloitte Access Economics, Employment Forecasts (June 2026 edition) — 82 “AI-disrupted” occupations expected to face the highest risk of declining employment, job vacancies in some already falling, and the blue-collar workforce declining 1.0 per cent or 37,100 workers in 2025–26; section 2.
12. Food Standards Australia New Zealand, Standard 4.2.5 — Primary Production and Processing Standard for Eggs and Egg Products and the P1060 Decision Regulation Impact Statement (September 2025); Egg Farmers of Australia policy statements; and Agriculture Victoria — exemption from accreditation and stamping for producers under 50 layers or 240 eggs a week selling direct in New South Wales, Victoria, South Australia, Western Australia and Tasmania, with Tasmania limited to under 20 dozen a week to friends, family and colleagues; the full standard above that threshold requiring accreditation, a unique stamp on each egg, traceability, auditing and Salmonella control; Egg Farmers of Australia stating it will request that state exemptions not be permitted and that all eggs be stamped; and mandatory Victorian poultry welfare regulations from the first half of 2026 applying to all poultry owners including backyard keepers; section 4.
13. Petrie-Flom Center, Harvard Law School, Cottage food and food freedom laws (LawAtlas data); ShelfLifePro, Cottage Food Laws 2026; and state cottage food guidance — every United States state having a cottage food law, with Arizona the last to legislate in 2022; approximately $2 billion in national revenue in 2024 from a negligible base a decade earlier and registered producers roughly doubling since 2020; permitted categories limited to non-potentially-hazardous shelf-stable foods with a pH below 4.6 for pickled goods; mandatory labelling of ingredients, allergens, net weight and a home-production disclosure; annual revenue caps between $5,000 and $75,000 depending on the state; approximately 35 states permitting pickled products and around 20 permitting acidified foods; and Wyoming, North Dakota and others operating broader “food freedom” laws extending to meat, dairy and prepared meals; section 5.
14. HM Revenue and Customs trading allowance guidance, and Simply Business with the Federation of Small Businesses — a £1,000 tax-free trading allowance introduced in 2017 requiring no registration, return or notification below the threshold; a separate £7,500 Rent a Room allowance; almost two-thirds of young people wanting to become entrepreneurs against 16 per cent actually starting; 28 per cent of young people having a side venture, rising to 40 per cent in London; and the recommendation to double the allowance to £2,000 and rebrand it a Side Hustle Allowance; section 5.
15. The Conversation, Small businesses in Australia have a hard time getting finance (July 2026), drawing on APRA and Reserve Bank of Australia data — approximately 50 per cent of small business loans backed by residential property, a further 45 per cent by other collateral and fewer than 5 per cent unsecured; lending to small business growing by around $58 billion against roughly ten times that in housing; non-bank lenders nearly doubling their share of smaller business lending since 2019 to around 27 per cent in 2025; the British Business Bank guarantee enabling lending that 51 per cent of borrowers would not otherwise have received; Germany’s KfW lending alongside commercial banks to share risk; France’s Bpifrance providing loan guarantees; and an OECD review finding public credit guarantee schemes increased small business lending while reducing collateral requirements; sections 1, 5 and 7.
16. ScotPac SME Growth Index, and ScaleSuite, Australian Small Business Loan Approval Rates and Lending Statistics (2026) — a record 47 per cent of small and medium businesses preferring non-bank borrowing; approval timelines of 24 to 72 hours with fintech lenders against 21 to 35 days with major banks, and around 80 per cent of non-bank borrowers funded within one week; an average small business loan of approximately $147,000, rising to $168,000 in Western Australia; and insufficient collateral and trading history under two years among the most common reasons for rejection; section 7.
17. Corporations Act 2001 (Cth), Part 6D.3A, and Birchal — crowd-sourced funding introduced in Australia in 2017, permitting eligible companies to raise up to $5 million a year through a licensed intermediary; section 7.
18. Productivity Commission, Annual productivity bulletin 2026 — agriculture, forestry and fishing leading all Australian industries with multifactor productivity growth of 10.4 per cent over the year, against a national decline of 0.5 per cent; section 6.
19. Australian Institute of Business (June 2026), LinkedIn Jobs on the Rise 2026, and Flowtivity analysis — 80 per cent of Australian small businesses using or planning to adopt AI; adoption rates of 69 per cent in property services, 65 per cent in finance and insurance and 61 per cent in business services; AI literacy as the most in-demand skill in Australia; and a median salary of $143,000 for workers with AI competencies, a 56 per cent premium over the workforce median; section 6.
20. Prime Minister of Australia, AI in Australia’s Interests (speech, University of Sydney, 15 July 2026), and White & Case, Australian AI Update — establishment of an Office of AI within the Department of the Prime Minister and Cabinet on 15 July 2026 to coordinate AI policy and develop Australian Standards for AI; legislation expected in early 2027 covering large-scale data centres, AI training and the use of Australian copyright material; obligations on data centre operators to underwrite new power supply without passing costs to consumers, to pay for grid connections and water infrastructure and to minimise water use; and the position that Australian books, music, art and news should not be used to train AI without the creator’s consent; section 6.
21. Department of Employment and Workplace Relations, Self-Employment Assistance and Free TAFE programme materials, and the Free TAFE Act 2025 — Self-Employment Assistance replacing the New Enterprise Incentive Scheme from 1 July 2022, providing twelve months of free mentoring, workshops and accredited training through providers in 51 regions, with a Self-Employment Allowance payable for up to 39 weeks, rental assistance for up to 26 weeks, and training treated as an approved activity for up to 13 weeks; accredited qualifications at Certificate III and IV in Entrepreneurship and New Business; the Fee-Free TAFE Skills Agreement delivering $1.5 billion for over 500,000 places from 2023 to 2026; and over $1.6 billion committed to 2034–35 for at least 100,000 places a year from 2027, prioritised to First Nations Australians, people aged 17 to 24, people out of work or on income support, unpaid carers, women facing economic insecurity and people with disability, in the national priority fields of care, construction, technology and digital, and the VET workforce; section 8.
22. Sovereign Australia, The Australian New Deal federal platform, Part 5.4, and the party’s published small business, cannabis, tobacco and cost-of-living policies; sections 6, 7, 9 and 11.