Subject: Re: Minting FS credits for value creation
From: Jamie Lokier <jamie@shareable.org>
Date: Sun, 8 Feb 2004 01:59:36 +0000

Forrest J. Cavalier III wrote:
> In other words, there is way too much value
> already being created in Free Software without
> hard currency rewards, that it might be silly
> to think we can foster much more with hard
> currency rewards.  

Speaking only for myself, my desire for hard currency in exchange for
writing Free Software is mostly to pay for shelter, food, transport
and taxes.  It would be nice to develop an economy where those things
can be had in more direct exchange for the development of the software
- removing the need for much hard currency at all.

> There is a big advantage in soft credits though:
> We can mint as many soft credits as we want, as easily
> as we create a database record.  If the signals of transactions
> are still present, relative value will become apparent, and
> people can and will use the accumulation of soft credits
> to guide the flow of hard currency (outside the system) towards
> popular features and successful projects and developers.

Yes.

> Obviously, it is important to keep even the soft
> credits scarce, because that is the only way that
> people will be careful with spending them.

Not flooding the system with minted credits is important.  But
"scarce" is not necessarily a good goal either.  If soft credits are
scarce, people will save them and avoid spending them on "minor"
projects.  If soft credits are minted regularly, people will spend
them more quickly precisely because of inflation, and be willing to
put them towards more projects, even frivolous ones, and be keener to
work at current rates (i.e. follow current valuation trends) than to
live off accumulated savings (i.e. not care about what people are
valuing at the moment).

While this increases the "waste", because even the most frivolous or
small projects receive credits easily, it is not really waste because
the relative signals, which is the purpose of these credits, are still
present, perhaps more so, when people are keen to spend.

It is not good if the vast bulk of payment signals are seen only by
the few most prominent projects.  That overvalues network effects,
i.e. the ability to market projects becomes more important than
developing the projects and letting money be exchange in ordinary
social relationships

> First, I think it is important that we do not mint
> any credits unless someone increases the size of the
> pie.  This is to stave off inflation.

The important thing is that the relative credit signals are present,
and the assignment of values accurately reflects the principles we are
striving to realise.

[ If you wanted to get funky, you could _define_ a relative valuation
of soft currency, because you know exactly how many are minted at all
times.  So when the Apache project has collected 10^6 "softs", and you
know that exactly 10^9 were minted to date (computed in real time),
the Apache project's valuation is 0.1% of the whole pie.  Multiply by
a million (fixed constant) just to make the numbers look
comprehensible, and you'd say that Apache is worth 1000 in the Soft
market.  _That_ is the value indicator to advertise, and, because we
have computers to help, it is the value indicator used when
negotiating transactions, and exchanging with hard currency. ]

> Hopefully they are appropriate controls so that no-one
> can game the system and let inflation sneak in....

Inflation is useful to disperse stasis.  Consider in the UK the
housing market is overpriced from the point of view of first time
buyers, most of whom cannot afford a house now.  House prices do not
fall rapidly unless there is a catastrophe; instead, people choose to
wait longer to sell.  So, for first time buyes, inflation is good
because they need incomes to rise until the reach a level where houses
are affordable again.  In other words, inflation devalues houses but
in a way which is acceptable to owners and banks.

I think the most effective way to prevent inflation is to _define_ the
money to always be a fixed size pie for the purpose of representing it
in valuations and negotiations.  Minting devalues the existing soft
currency, but that's ok because that encourages circulation.  It is
similar in some effects to paying everyone a fixed income every week,
to spend as they wish (you could do that as well, as it is a nice feel
good motivator).  The rate of minting needs to be a democratic policy
decision and reviewed regularly.

> Paid or otherwise limited membership is probably going to be
> important to keep out the rif-raff, but maybe one paid
> membership could come with 3 free-invites for your
> friends/colleagues/whoever, or maybe 3 project insertions.
> Or maybe you get one invite a month.

Just let people pay a hard currency membership fee.  It seems to work
very well for elance.com, where the quality of work offered and
professionalism of the contractors who do it is much higher than the
free sites such as CoSource and the like.

> How do we have more ways to earn credits, but not cause inflation?

I must ask why do you think it is important to avoid inflation?

We have all read horror stories of hyperinflation and empty
supermarket shelves, long bread lines and such.  Those are real
stories.  But there are also stories of high inflation where, despite
rising prices, people seemed to have enough money to buy what they
needed.  That strikes me as a desirable situation.

> I think we can decide that 1 credit should be equal to
> a nominal amount of someone's time.  More highly-skilled
> people will be able to work faster or choose or be able
> to complete the closure of DB records with higher assigned
> credits.
> 
> I am nearly certain we do not want to just mint new credits for
> each member to spend every day.

If you have members paying a hard currency subscription, then minting
new credits for them every day will encourage that (people are getting
something in exchange for their subscription), and lets members feel
good that they always have some power in exchange for their
involvement even if they aren't produces of code themselves.

Paying everyone daily devalues accumulated net worth of large projects
and popular developers, so that for something/someone to _retain_ a
high net worth, they have to continue being socially valuable enough
that people keep paying them.  In other words they need to keep being
relevant.  When everyone is receiving new credits daily, they will be
more inclined to keep paying those projects they value - keeping the
valuation system in equilibrium.

It would be good to automate daily payments from a person to a
project, too, allowing them to automate a regular contribution x% of
their income to the GNU Frob project for example, with the ability to
change their automated contributions at any time.  In effect it allows
people to contribute value when they want, but to take it back and put
it somewhere else when they want.

In some ways that's a _better_ value indicator than accumulated
wealth, due to the smaller and faster information feedback loop.  It
is more like direct democracy than representative democracy.

[ Warning: That means it has the problems of direct democracy too.
  Misinformation, marketing, advertising, offensive business practices
  that you don't hear much about etc. could easily play a significant
  role in people's spending patterns.  This is not good - it is
  important to create a system in which real information, and real
  involvement is encouraged, not one which encourages people to sit at
  home basing all ther allocation decisions on hearsay and spin from
  popular news sources and glossy brochures. ]

In a system like that, you _want_ inflation.  People receive credits,
but you want to encourage them to spend them too, which inflation
does.

Summary: Mechanisms to promote more rapid information from creditors
to workers, restrict memory in the system to prevent accumulation of
relatively vast wealth from past successes, so that work and focus of
work stays relevant to creditors' update signals, and creditors are
encouraged to keep those signals maintained.

> I think ExpertsExchange and similar sites did this to jump-start
> participation.  Isn't that a quick way to cause inflation?

It is, but what is the problem?  I'm sure I'm not seeing what it is,
so could you explain why it is important to avoid inflation,
especially given that the amount of credits in circulation at any time
can be used to compute relative net worth in real time?

>    Can we tie it to lines of code?
> 
>    Maybe it should just be 1 point for every 100
>    lines of documentation.  :-)  Get docked points
>    for having an unreliable project website.
> 
>    Official maintainers can assign 1 point for each
>    10 lines of code they accept to close a DB record,
>    or the amount that a DB record collected.
> 
>    1 point for submitting a bug report with a patch.
>    
>    1 point for reviewing/trying a patch.
>
> Mint points for participation on mailing lists?
> 
>    1 point for each post to a project message board,
>      awarded after a delay.
> 

Every one of these encourages people to game the system, because of
the virtual financial reward for every individual step.  They don't
encourage community building, which is the bedrock of quality work.

>    Projects come in with 1 point for every 100 lines
>    of source code or documentation.  But this should
>    be amortized or somehow vetted by the community.

Vetted by the community is better, with danger of guilds forming and
consolidating - newbies getting excluded.

>    -2 points if "enough" others reclassify your post during
>     URL review as noise, (E.g. "ME too.", "send me a point")

Now we're into problems that Slashdot and their ilk have, of trying to
use democracy to classify quality contributions, yet protect against
cliques of people trying to break the system.  There needs to be an
incentive for people to not do that.

> One other advantage of soft credits is that the system becomes
> easy enough to set up for one project: there is no need to
> have a central website or middleman bank: 

Are you suggesting a system based on a form of digital cash?

> This would enable an individual project to set it up themselves,
> and do the work to attract members. They would then have a way
> to let the community do distributed decision making.  
>
> Once a few of these sites were set up, if they used the same
> way of assigning value, the credits would be compatible, and
> you could transfer credits between sites.

Until there's a large consensus on methods, there will be competing
sites with different methodologies.  Therefore consensus building is
important.

To the extent that different sites with different methodologies will
exist, it would be good to have some kind of exchange system.  Given
that an exchange with hard currency is also desired, it seems quite
feasible to include exchange with soft currencies too.  The important
thing here is to put in place the motivations which reduce gaming the
whole system of currencies.  That means non-technical stuff: community
building, projects driven by more than just economic motivators.

> (Well, it would be a mess, because you have to stop counterfieting.
> Maybe transfer credits between sites should only be with hard
> currency and gentlemen's agreements.  Kind of like selling your RPG
> account on ebay.  Or maybe a parent project like collab or the ISC
> could control several cooperating sites and ensure they didn't cheat
> with federal funds.)  Or maybe it is more like corporate mergers and
> stock swaps....

You can fight counterfieting - or you can just allow it.  So what,
someone gets rich one day, they get found out, and the normalised
inflationary system I describe earlier depletes their riches rapidly
because people don't trust that person any more.

(Trust networks are another issue.  I'm not convinced that anonymous
currency transactions are ideal for this).

More and more I am convinced we should all read up on the Mondragon
cooperatives, which are people doing lots of real work and
economically very succesful, to see how their funding system works.

It isn't perfect, in fact it seems to be starting to lose track of its
core principles which make it work lately, but it's remarkeably
successful on a large scale and well worth learning about.

-- Jamie