Skip to content
Descriptive findings · 4-minute read

Can corporate tax fund a world with less work?

The available data help frame the question, but do not yet provide a causal answer or guarantee tax revenue.

MacroVision editorial team · Updated:

In one minute

A statistical association between tax rates and revenue does not establish how much a future tax increase would raise.

Candidate events201
Complete observation window166
Isolated events25
Corporate-tax-rate changes screened in step 61: at least one percentage point, with a 3-year window on each side. Isolated events do not establish causality.

What has been tested?

Step 59 compares panel specifications across 38 countries and 905 overlapping country-year observations, using current and prior-year tax rates with growth and inflation. Existing research records, not a live update.

What can we learn?

When current and lagged tax rates are included, the current-rate coefficient is about 0.0504 percentage points of corporate-tax revenue relative to output for each percentage-point increase in the statutory rate. Its reported confidence interval excludes zero; the lagged-rate interval includes zero. The result warrants further testing, but does not establish causation.

Why not use this coefficient directly in the budget calculator?

Governments may adjust taxes in response to economic conditions. Profitability, tax-base reforms and international profit shifting may change at the same time. Adding a time lag does not solve the causal-identification problem.

What are the limits of the research?

Step 61 screened 201 tax-rate changes and found 25 isolated candidates. No event-study effect or fully reviewed control assumptions. Step 62 examined before/after profiles: five tax increases averaged about 0.0573 percentage points of output in revenue change one year later, but leave-one-out means range from −0.0800 to 0.3299. Sensitivity range, not a confidence interval; no control group in step 62. Step 63's stricter stability/support filters retain 6 events, including two increases. Descriptive feasibility only; control candidates lack institutional approval and 19 events await source review. Step 64 finds accompanying policy packages and anticipation risk in the two retained increases; comparison countries are not clean controls. Step 65 downloaded 12,653 Korean company-years, allowing a descriptive pilot but not a verified actual tax base for reform exposure. Step 66 found accounting reconstruction gaps and paused the pilot. Step 67 reconstructed tax ratios in two original separate-statement cases and rejected treating tax as cash paid. Step 68 expanded to six observations: prior-year inputs and numeric precision explain many gaps, but exceptions and processing definitions remain unverified. Step 69 reconstructed remaining leverage/size exceptions using observed bounds: a post-hoc numeric explanation, not independent pipeline validation. Statement checks reached nine observations across three companies. Step 70 completed a limited public-statement pilot showing tax-expense/cash-payment differences and weight sensitivity. Fiscal calibration remains blocked.

Sources and transparency

This publication explains the existing research in accessible language. Its accounting example and scenarios are neither official forecasts nor investment recommendations.

Review data and research limitations ←

What do you think?

Which benefit-funding risk matters most: a shrinking tax base, profit shifting or slower growth?

Further reading

Data guide · AI and work

Which occupations are most exposed to AI? The full ILO ranking

Office and clerical work at the top of the list, physical work at the bottom. What the International Labour Organization's index really measures, and what it does not say about layoffs.

Empirical finding · AI and work

Is AI already eliminating jobs? What the data from Europe show

We examined 13 countries and 39 occupational groups: did exposed occupations lose more jobs in places where firms adopted AI faster? The answer for now: no effect can be detected.

Research finding · Simulation

Even when benefits protect them, those who lost their jobs fall behind

In all 42 matched comparisons in the automation simulation, displaced workers ended below the path in which the change did not happen, even when benefits maintained their level of consumption.

Research scenario

If AI replaces jobs, what will we live on?

An unemployment rate of 20% raises a question beyond the number of jobs: how are the gains from automation distributed?

Accounting illustration

Prices have fallen. Will the benefit buy more?

Purchasing power also depends on wages, taxes and debt. A simple example shows why lower prices are not the whole story.

Descriptive case study

Tax expense and cash tax — why do the numbers differ?

A case study of three companies illustrates the difference between tax expense and actual payments, and the effect of weighting choices.

Scenario analysis

20% global unemployment: an inevitable crisis?

An extreme scenario can lead to very different outcomes — depending on income distribution, debt and public financing.

Economic explainer

Cheaper machines, cheaper products?

The link between automation and prices depends on competition, non-labor costs and how quickly the economy adjusts.

Concept guide

Could a basic benefit preserve living standards?

A purchasing power test must include prices, debts, taxation and the source of the benefit's financing.

Distributional analysis

Asset ownership versus debt: who is more resilient to automation?

The difference between labor income and capital income may become decisive — but asset owners are also exposed to risk.

Behind the research

Behind the forecast: how do you build a model for the age of automation?

From scenario assumptions through government accounting to sensitivity tests — and what still cannot be concluded.

Research finding · Simulation

Falling prices do not reach everyone: why renters are hurt more

In the automation simulation, when savings pass through to prices, renters and mortgage holders fall behind. Indexing the benefit to housing costs helps only when there is budget.

Follow the research

You can also follow new publications via an RSS feed, without leaving any details.

Open publication feed ←

Basic site measurement counts page requests and reaching the article end after reading time. No measurement cookies, personal identifier or IP address in measurement records. Counts are not unique readers.