About Digital agent income (and some implications)


There are three broad classes of income :

  • Earned Income: is generated by working, such as salary or money made from time-based employment which can include consulting, working a job, running a small business, gambling, etc. Income is linear with the activity: once you stop working, you stop earning.
  • Portfolio Income: is generated by selling an investment at a higher price than the original cost (capital gain). As life is, these investments  include a high level of risk as one can generate gains (generally) proportional to the capital invested (that can be reinvested increasing returns) or losses Having a crystal ball could make it grow exponentially. Income is not linear with the activity.
  • Passive Income: generated from assets purchased or created. Examples can be income from rental of real estate or other form of invested capital, business income (not based on amount of time and effort spent), income from intellectual property sale. May be (and generally is) recurring, generally linear with the assets employed.

All of these incomes historically could be statically attributed to a specific limited number of contracting states (a good or service being produced in one state and sold in another). The OECD has defined a convention on income and taxation which further details classes of income and the way they should be treated when operating across borders. These are:

  • Income from immovable property
  • Business profits
  • International shipping and air transport
  • Associated enterprises
  • Dividends
  • Interest
  • Royalties
  • Capital gains
  • Income from employment
  • Directors’ fees
  • Entertainers and sportspersons
  • Pensions
  • Government service
  • Students
  • Capital

all of these classes nail down to the three broad classes described at the beginning.

if we think to the income generated by SAAS (software as a service),

  • it is non linear with the activity, time or effort;
  • it is not proportional to the capital invested,
  • it is recurring.


  • it may involve several, dinamically changing, contracting states (transactions online may be conducted in milliseconds between parties resident in a number of different states, under the umbrella of framework agreements)
  • it can grow exponentially (not employing labour, variable capital and not incurring in losses), and
  • it has (almost) zero marginal costs, returns which can be disproportionally higher than investments so these may be considered sunk investments

these last considerations make it a very different type of income compared with  types of income we have known so far. it’s income generated by a machine (like it would be for capital investments with variable costs and a production location), but (substantially) without the capital investment, without the variable cost, and without the production location.

it is not simply “automation” as this term is already used to describe situations where variable costs exist, where capital investments are important, where the income cannot grow exponentially, where contracting states are stable.

we could rather call it an Digital agent income

(it may, but doesn’t need to be “intelligent” like we mean for artificial intelligence / machine learning applications)

we could agree on specific taxation criteria, like for example a rate similar to capital gain, opt-in conditions, exemptions for startups, exemptions for investments, etc.


This is the basis of the bill proposal I made while I was in the italian parliament, a couple of years ago.


Now, let’s make a step forward:

as I argue in my latest books, the immaterial dimension has become the major user interface of the material dimension for socioeconomic relationship: “dimensions” and not “worlds” because they are not alternative to each other but rather they supplement each other; material/immaterial rather than real/virtual because they are both real and because the world “virtual” (which comes from medieval latin ‘virthualis’) has a root meaning of “potential, not real”.

we tend to think at digital technologies as enablers. and it certainly was so, when we bought hardware and software for specific applications. the move to SAAS ‘platform’ business models (multi-sided markets) has given birth to intermediaries that capture value monetizing the relationships between the intermediated. often leading to mono/oligo-polistic/psonistic powers exploiting network effects and lock-ins given the absence of pro-competitive regulation imposing portability, interconnection and interoperability.

they are among the few being able to capture a portion of the monetary value that is destroyed by digitalization, as the value-carrying portion of goods and services moves from material to immaterial and hence to zero variable costs. (which implies also that their prices are not determined by costs but by perceived value).

I believe this has a deflationary effect and induces a pressure on earned  income (and, less significantly and in a longer term, on some forms of passive income), at least expressed in monetary terms, increasing polarization. It does not necessarily imply that persons’ well-being reduces, as many goods and services benefit form a reduction of prices due to the same reasons: prices of goods and services being increasingly impacted by migration of value-carrying parts to zero variable costs.

with the waves of industrial revolutions, generated surpluses first moved peasants to industry and the same happened for the secondary sector. The second industrial revolution gave birth to the tertiary sector and the third industrial revolution to the quaternary sector (or advanced tertiary).

I am inclined to think that the surplus generated by artificial agents is both, having the effect of a reduction of the wealth expressed in monetary terms and a significant portion of it being captured by digital intermediaries in the process of progressive de-materialization of significant parts of the economy, contributing to the increase in polarization.

Someday, maybe we will be able to store energy generated by renewable sources, further increasing the portion of the value of the economy that is linked to the dynamics of zero variable costs. There may be interesting times ahead.

If you like this post, please consider sharing it.

1 thought on “About Digital agent income (and some implications)”

  1. Stefano Quintarelli

    One could envision non-deductibility of part of the VAT on Digital Agent Income. Given the peculiar characteristics of this type of income, the effect would be to drag down the price of the auctions, reducing the supplier’s revenues by a corresponding amount, without noticeable effects on consumption and costs transferred to end users.
    It could even be an extraordinary temporary measure, related to the Covid-19 emergency, favoring an agreement on a new type of income at the OECD.

Leave a Comment

Your email address will not be published. Required fields are marked *