I am a big proponent of paying interest for real money prediction markets, as Kalshi and IBKR do.

But in mana markets (not sweeps/real money), the interest rates you'd need to pay to make this work (assuming it's a replacement for loans) are really high, probably like 5% a month given that's roughly the going rate for peer to peer loans, i.e. that's a good estimate of the discount rate.

Loans solve the problem a different way, by reducing the opportunity cost, by reducing the amount of capital you lock up.

I think interest rate mechanisms are easier to understand and more robust, whereas with loans you have to worry about people going bankrupt etc. But paying that much interest would dramatically change the mana economy, I'm not sure what that impact would look like and whether it would be bad but that is in my mind the key difference between doing it in mana vs real money markets.

@jack I think even a small interest of a few % per year which is paid on the invested mana, but not on uninvested mana would be a good incentive. Like, if someone wants to make long-term predictions but not daytrading, it would make it more viable.

@OlegEterevsky Resolves Yes given loans? I believe they count for the market description here, which is "Will Manifold implement any kind of growth or interest mechanism for invested mana?", allowing you to get a form of interest on your investments (that you have to pay back but I think it counts as "any kind")

@Fay42 i don't personally think that the site giving you a loan (but practically what is happening and also explicitly what the site calls it) is equivalent to paying you interest. those are somewhat opposing ideas? compensating you the cost of locking up your capital vs freeing up that same capital for other uses.

if this question was solely about "incentivizing betting on long term markets", then absolutely loans would count (this discussion was prompted by the removal of loans). but the title is pretty specific that this is about "automatic growth of invested mana"—giving someone a loan is not the same as paying them interest (or growing their investment).

I think it's a kind of interest mechanism, even if it's not the one envisioned? I do think it's not clear either way.

@Fay42 @Ziddletwix Could you guys link the announcement? I think I missed it.

FWIW, I don't think any kind of loan mechanism qualifies as invested mana growth. All of the linked comments describe a fairly idea, which doesn't involve loans.

@OlegEterevsky I don't think there was an announcement on the site—in Discord they were announced like this:

(it works basically like loans previously did, but at a slower rate)

Bit that seems relevant is the size of the loan is determined by your investment - you get a percent of your invested mana (minus previous loan amount).

One option would be to simply make this an alternate way of adding liquidity. That is, users could subsidise a long-term question by investing some mana to be paid out over time to those invested in it.

Manifold currently prints mana for all users. It also wants to guarantee cash withdrawals based on that mana. I think people underestimate how costly it would be for manifold to offer interest on all of the mana it prints, compared to the typical status quo, where betting sites can help fund their operations with 0 interest on all deposits (typically not including “deposits no customer paid for”).

Simple thought experiment: Manifold currently closely tracks how much mana it prints each day. It views it as a serious/substantial cost. Compare the amount of mana currently printed each day, with the amount that would be printed from interest on all invested mana https://manifold.markets/stats. This is not a trivial hit to Manifold's margin, it would be an enormous change.

Conversely the current state of affairs means that the bets are much smaller than they could be.

Consider this question, for example. It's subsidized by 10k mana. Theoretically it means that it should attract more traders than another question subsidized by just 1k mana, since it provides more returns for the time you spend thinking about it. But the current rules disincentivize substantial bets for the events a year from now and this question so far has attracted only a few hundred in stakes. Paying interest on the investments would make bigger bets much more viable.

Of course, Manifold has to somehow pay its bills, but I assume it could operate the same way as the real-life stock market brokers by taking operation and custody fees.

That's fair! Interest on the mana that Manifold prints is expensive. Paying interest on $ deposits is expensive, relative to the status quo. But if $ deposits massively increased because they paid a bit of interest, & Manifold was able to sufficiently scale up so they are still earning enough of a margin on all deposits (paying users less interest than they receive, taking enough fees from withdrawal, etc), that's a credible theory of how they can accomplish this.

But it requires fairly strong assumptions. And I'm extremely skeptical that manifold could pay "interest on deposits" and avoid being regulated like other financial institutions that do this.

As long as Manifold operates on play money, it can do whatever it wants to. As soon as it switches to real money, it will be in same position as Polymarket, regardless of whether they pay interest or not.

Polymarket is “not operating in the US”, and is very much not following US regulations. Manifold could try that, but it’d be an enormous change for the site/employees

That's exactly what I meant. As soon as Manifold decides to lean into real money markets, it will have to follow the steps of Polymarket. At that point paying interest won't make matters worse.

I think this happens eventually, but not until they reach product market fit and raise a successful series A. It's a super helpful but not core functionality and would take a lot of time to make it work.

Since Manifold currently operates on play money it shouldn't be that difficult to implement.

I think we've reached the point you can't say Manifold runs on play money anymore. Prize points are here. Manifold cant treat mana as expendable like they previously could.

Prize points constitute relatively small portion of Manifold "assets" and from my understanding can't be directly invested, so wouldn't grow with this mechanism.

There are exploits to turn mana into prize points, so I don't think you can view them as substantially different. I think the bottom line is that implementing automatic growth of mana invested would likely mean coordinating real dollars invested in some bond or index fund to back it up. Otherwise it's exploitable.

+1 up until this point, it has been trivial to turn large amounts of mana into prize points (at the minimum, in expectation, but at least currently, also literally, because of multiple choice questions with impossible answers). for the purposes of accounting, they have to be treated the same