NextLevelCorporate (R)

View Original

AI’s great, but who pays the tax?

Image: Pavel Danilyuk

TL;DR

Artificial intelligence (AI) is increasingly being used across various sectors, including healthcare, finance, transportation, and entertainment.

While it has led to greater efficiency, productivity, and innovation, it also poses significant challenges to governments, particularly in terms of tax policy.

Tax receipts are the main source of revenue for governments and are used to fund public services and pay government employees.

But, if AI technologies are used to replace human labour, leading to higher unemployment and lower demand in the economy, who will pay the tax?

Governments may need to rethink their tax systems.

Governments could consider moving closer to a consumption tax. The adoption of a consumption tax might be less harmful to the labour market than an income tax as it is based on spending rather than income. But a consumption tax can be regressive, meaning that it can disproportionately affect low-income households.

Other alternatives? Impose an AI super tax on companies that provide (or use?) labour-disruptive AI, or a corporate tax rate increase across the board given pretty soon every business will be using it, or issue more government bonds, reduce government size, spending and budget deficit, or perhaps something else.

So, who will pay the tax? The answer is that you will.

And here’s why 👇

Pervasive AI

AI is becoming increasingly pervasive in our society, with applications in a wide range of industries and sectors.

In the healthcare industry, it is being used to develop new treatments, improve patient outcomes, and enhance medical research. AI-powered tools can analyse large volumes of patient data to identify patterns and predict outcomes, helping healthcare providers make more informed decisions about patient care.

For example, AI is being used to develop personalised treatment plans for cancer patients based on their genetic profiles, improving the efficacy and safety of treatments.

In finance, it is being used to detect fraud, assess credit risk, and develop investment strategies. AI-powered algorithms can analyse shed loads of financial data in real-time, identifying trends and anomalies that human analysts may miss.

This enables financial institutions to make more informed decisions about lending and investing, reducing risk and improving returns. Think about FTX….

In transportation, it is being used to develop autonomous vehicles and improve traffic flow. AI-powered algorithms can analyse traffic patterns and predict congestion, enabling transportation planners to develop more efficient routes and schedules.

Additionally, it is being used to develop self-driving cars, trucks, and drones, which have the potential to revolutionise transportation and logistics.

With the increasing prevalence of ChatGPT and other AI in these and other industries (yup, thousands of competitors are springing up to chase VC funding) it’s clear that AI is becoming a ubiquitous part of our lives and its impact is likely to continue growing in the years to come.

Still, the adoption of AI also poses significant challenges to governments, particularly in terms of tax policy.

Tax receipts are a critical source of Government income

Tax receipts are important to a government as they represent a significant portion of the revenue needed to fund public services such as education, healthcare, infrastructure, and defence.

Tax revenue is also used to pay salaries of government employees and service government debt.

Take Australia for example. Here is how the Government is funded, broadly.


Personal taxes make up 48% of government funding and company tax ~18%. GST makes up around 13%.

Think about that for a minute - the haul from our form of consumption tax (10% GST) is nearly as large as from corporate tax (28% to 30% tax rate).

This is because other than for exempt bodies, all companies, trusts and individuals pay 10% GST (less input tax credits where registered for GST) whereas many companies pay effective tax rates well below the corporate tax rate due to generous tax deductions, structures and domiciles.

Without tax revenue, governments would be unable to fund essential services, which could lead to social unrest, political instability, and economic strangulation/stagnation - unless treasurers were to glue their elbows to the ‘let’s just borrow more’ button on their desks - none of which are desirable outcomes.

So, if AI technologies are primarily used to replace human labour, and this leads to higher unemployment and lower overall demand in the economy, then a government that is supportive of AI but also wants to preserve its tax receipt income might need to so something different.

How to solve the tax gap

Here are some choices:

  1. Ban AI (really?)

  2. Move closer to a consumption tax

  3. Impose an AI tax on companies that provide (or use?) labour-disruptive AI, or a corporate tax rate increase across the board

  4. A mix of 2 and 3

  5. Reduce government size, spending and budget deficit (now there’s a thought 😉)

  6. Issue more Government bonds to fill the gap

Let me briefly interrupt the thoughts of any protectionist readers by saying that I will not be considering choice 1 today 😁 and while choice 5 is logical, it’s a pipe dream.

Being based on spending rather than income, a consumption tax could still generate revenue from the consumption of goods and services while not discouraging work and investment as some jobs are automated, but low-income households might bear a disproportionate burden of the tax.

That’s why it would probably need to be a progressive consumption tax and/or provide exemptions for basic necessities like food and housing.

As AI technologies become more prevalent, we might see significantly higher levels of productivity and efficiency, which could in turn lead to higher profits for companies. This could potentially result in higher corporate tax receipts for governments and returns for shareholders.

But displaced income taxes make up a way greater share of the tax haul than corporate taxes. Levying a tax on specific AI-related activities might also be an option, but difficult to track and enforce.

A mix of a lower income tax rate for all, paired with higher (yet progressive) consumption taxes, plus a modified corporate tax rate might work, however engineering a mix that would be palatable to governments and communities is highly unlikely.

So, what are we left with?

The most obvious answer is that governments will need to issue more bonds to fund themselves (and welfare) to replace income taxes, and that in turn will add to the debt burden.

The logical conclusion here is that interest rates would need to fall simply to allow sovereigns to fund the servicing cost on the additional debt - and in any case that’s where we are now, and in some countries, we’re still quibbling over whether a government can raise its debt ceiling to continue functioning.

So, while AI may be an inflationary force in terms of increasing productivity (if that is what really happens) it may equally become a deflationary force, or a tax on all of us if governments need to issue higher levels of debt to fund themselves in a world where income tax receipts collapse due to AI driven labour displacement.

AI becomes a tax if we all end up paying for it in the form of less purchasing power as a result of government treasuries having to fill the gap with new debt issuance.

That’s also why QE Infinity is a tax - it achieves its objectives by debasing your money and purchasing power as reserves resulting from central bank bond buying find their way into bank deposits which are then lent out (money creation) at cheaper rates (risk pricing).

For AI to not be deflationary would require the rate of real productivity gains from AI to outweigh the rate of money creation resulting from AI-displaced personal income taxes.

One way or another, you and I will end up paying the AI tax - because there’s nothing surer than debt, debasement and taxes.

See you in the market.

Mike

Image: Cottonbro