Does this line of reasoning make sense in terms of economic agents, even if implementation is highly unlikely?

  1. The Gini coefficient is an imperfect but reasonable measure of income inequality within a country.

  2. One reason that governments avoid increasing income tax on high earners is that they will move to another country with lower taxes.

  3. Therefore, if all countries agreed to a certain level of high-income tax ("we will all tax our highest earners at least 45%"), there would be no incentive to move and all countries could tax their high earners more.

  4. Governments could use the taxes to pay for labor, or basic income, which should improve the Gini coefficient for income in their country.


  • I've never quite understood claim 2, because firstly you are taxed on income not wealth, and secondly I assume you can't simply transfer huge amounts of money between countries. I've focused on income.

  • Although claim 3 seems unlikely, in 2021 the G7 reached an agreement on taxing corporations:

The United States, Britain and other large, rich nations reached a landmark deal on Saturday to squeeze more money out of multinational companies such as Amazon and Google and reduce their incentive to shift profits to low-tax offshore havens.



2 Answers 2


No for rather peculiar reason.

Standard Gini coefficient is more or less unaffected by redistributive measures since standard Gini coefficient is calculated from market income. Hence, taxing and subsidizing poor does not affect Gini coefficient directly at all. Auten & Splinter (2023) recently published paper about it (although working version was already available for at least 2 years).

Gini can be affected indirectly because high taxes might encourage high income earners to work less, and thus reduce higher incomes somewhat. But non-linear tax system create poverty trap that might increase it somewhat as well. Nonetheless, due to the way how its measured redistribution cannot directly affect standard Gini.

However, such tax harmonization could be almost certainly used to raise taxes higher. An important limit to the tax schedule, under lets say Rawlsian Max-Min preferences or other preferences that lead to redistribution, is the tax elasticity of labor supply. The higher the tax elasticity the lower the optimal taxes (e.g. see Saez 2001). Competitive tax rates and options to move increase the tax elasticity and lead to lower optimum tax rates and vice versa.

Hence, you could almost definitely increase taxes and redistribute more (unless the option to move to low cost country does not affect these tax elasticity a lot). However, you should not expect that to change standard Gini income inequality coefficient significantly, since the standard Gini by design ignores effect of taxes and subsidies. It would affect after tax and transfer Gini, but that is not Gini that you typically see reported in media or even research until just very recently.

  • $\begingroup$ is there a better measure than Gini for income distribution by which I mean "actual money paid into workers' bank accounts", so it would include government-paid labor and maybe even basic income, and actually measure redistribution? $\endgroup$
    – user43089
    Oct 31, 2023 at 12:24
  • 1
    $\begingroup$ this gives me a lot of leads to follow up on, btw, thanks so much! $\endgroup$
    – user43089
    Oct 31, 2023 at 12:25
  • $\begingroup$ @LeeMcGee it is possible to calculate Gini after taxes and transfers are factored in. That is precisely what the Auten & Splinter paper is doing. However, while standard Gini is available for almost every country the post tax and transfer Gini is not widely available, save some exceptions like that paper that calculates it for US. However, if you search for it you cannot search for Gini because that will give you the standard Gini, you need to try to search explicitly for post tax and transfer Gini. Other alternatives are looking at capital vs labor income but that has its own problems since $\endgroup$
    – 1muflon1
    Oct 31, 2023 at 12:29
  • $\begingroup$ nowadays even workers often derive capital incomes (from stock options, pensions, retail investing) so it is not that representative. Best measure is to look at cumulative distribution functions and examine level of inequality by comparing various cumulative distribution functions, although it won't just give you a single number to look at it is quite comprehensive way to understand inequality everywhere in income (or if you are interested in wealth wealth) distribution but they are quite data intensive to create $\endgroup$
    – 1muflon1
    Oct 31, 2023 at 12:32
  • $\begingroup$ @1muflon1 Gini are often based on consumption rather than income and so do often take account of taxes and transfers. For example, in the details for the Gini index data on the World Bank website (data.worldbank.org/indicator/SI.POV.GINI) it says "Wherever possible, consumption has been used rather than income. Income distribution and Gini indexes for high-income economies are calculated directly from the Luxembourg Income Study database, using an estimation method consistent with that applied for developing countries." $\endgroup$
    – smcc
    Nov 3, 2023 at 17:09

There is an unforeseen (or maybe foreseen and undisclosed) consequence of a global tax rate for high income earners: Migration to more developed countries.

If you make $300K as a software developer and you are taxed 45%. Would you rather live in Austria, or in the Philippines? The level of services you receive from the government do not improve as the amount of taxes you pay do.

Thus, poorer countries should tax the high income earner less as it can offer them less. Otherwise, they'll migrate to countries with better infrastructure.

It's not logical, in my opinion, to study a global rule at the country level. While this might improve Gini at the country level, it'll increase inequality between poor and rich countries.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.