Thus far my two posts have detailed the reasons why I am highly skeptical of steem's long term sustainability. I'd like to provide some concrete numbers from blockchain analysis and propose a possible solution for generating revenue.

HOW THE BLOCKCHAIN WORKS

Below is a simplified diagram of how the newly minted steem is divvied up (to the best of my understanding).

There's actually no such thing as steem power in the system. There's only steem dollars, steem, and vests.

Vests represent a claim to redeem steem from the vested pool upon powering down. The ratio at which you can do so is called "steem_per_mvest" which you can find on steemd. It's calculated by taking total_vesting_fund_steem / (total_vesting_shares / 1,000,000).

In other words, steem per mvest is the amount of steem in the vested pool divided by the total number of mvests owned by all accounts. The "steem power" figure in your wallet is calculated on the fly using that conversion rate but it will change constantly.

Every block 30 steem are created (since liquidity rewards have gone away). 2.5 steem go into the pool and result in new vests (at the current steem per mvest ratio) paid out as rewards to curators, content contributors, and miners. 0.5 steem get converted into steem dollars (at the conversion rate set by witnesses) and paid to content contributors. 27 steem goes straight into the pool and results in neither new vests nor steem dollars.

Overall then 98.3% of new steem goes into the vested steem pool, 90% being straight vested steem and 8.3% being vested steem that also results in new vests. The other 1.7% gets converted into SMD rewards.

THE BEGINNING

Since the steem per mvest figure is calculated with vested steem and total vests, how was the ratio determined in the beginning before any steem was vested? The 27 steem were not being dumped into the pool during the first week.

As far as I can tell, at that time the default ratio was 1 steem per mvest. Currently the figure is about 289 steem per mvest. Vests have gotten almost 300x as expensive in terms of steem since the first week.

To give you an idea of what that means, the first guy to power up was named "faddy." He put 357 steem into vests on March 25th, the day after mining began. 357 steem has a current market value of $528. But because he vested it so early his vests are currently valued at $103,000, at least on paper.

The official Steemit account is valued at over $100 million because they secured ~80% of the mining rewards during the first week and powered up 264,662 steem at one steem per million vests, giving them 264,662 mvests. Every week they power down (so far a total of 5.2 million steem and currently at a rate of ~$1 million per week) the steem per mvest ratio will increase resulting in more steem being paid out.

If you want to project the steem per mvest figure just take y = 1 + 2x where x is the number of days since the 27 steem started being dumped into the pool. This is because it's steem per mvest = (first-week-vested-steem + daily-new-vested-steem * number-of-days) / first-week-total-mvests. 1 first-week-vested-steem = 1 mvest, so that reduces to 1 + (daily-new-vested-steem / first-week-vested-steem) * number-of-days.

I believe there were ~390,000 or so steem vested in the first week, and at 27 vested steem per block (777,600 per day) that reduces to 2, hence 1 + 2x.

It may seem more complicated since powering up and powering down changes the number of vested steem and total vests. However these actions are always done at the current steem per mvest ratio so the conversion rate is unchanged.

HOW MUCH OF THE NETWORK IS POWERING DOWN?

I wrote a script (heavily borrowed from user jonblack who was analyzing bots, see here) which took the top 500 steem power holders from the rich list (who have 97% of network vests) and grabbed the entire history of their power downs.

(Side note - if you want to write a script for this be aware that power downs are not blockchain transactions, they are virtual operations which you can access by taking an account name and using get_account_history)

I excluded the official Steemit account because of the top 500 they comprise 74% of power downs. The remaining top 499 make up 94% of network vests. Below are the results.

The graph spikes every week because for whatever reason a lot of the power downs happen around the same few days. You'll notice also that the peaks arch upwards as time goes on. This is largely because the steem per mvest figure is increasing every day so the same number of vests powered down result in progressively more steem.

To get a feel for how much powering down there is across the network I looked at the last two weeks and how much there was from the top 499 vs their ownership at the end of that two weeks. It's not a perfect way to calculate it but it showed that they powered down about 55% of what they could have given the ~1% per week limit.

Looking at the bottom 450 out of that 499 they powered down about 43% of what they could have in that same two weeks.

So overall the network is powering down more than I expected, somewhere between 40 - 50%.

HOW MUCH OF THE NETWORK IS POWERING UP?

I also ran a script looking through blockchain transactions for power ups from July 4th (the day of the first payout) through August 19th. The power ups are in blue and the power downs are in orange (the official Steemit account is excluded again here as they did not power up in that time).

I also made another chart to show net power ups (power ups minus power downs) for the same time frame.

I was somewhat surprised at how much powering up there was, but also you can see the chart trends downward. Again this is largely because the power downs are resulting in progressively more steem every week for each mvest.

ANALYSIS

The main takeaway I think is that assuming the powering down remains at 40-50% for some time the effects of steem inflation will become more pronounced each week. Given that currently about 4% of the total steem supply is in liquid form, the steem price is much higher than would otherwise be the case if it was all tradeable.

The top 49 whales from above have 66% of the total vests. At some point their powering down might slow as they're no longer willing to trade voting power for steem. But this will take a while given the weekly limit and the fact that many of them are reaping large curation rewards. As such the new steem on the market each week might cause a long term down trend on the price.

As for power ups, I don't know how much of it is new investment and how much is steem dollar rewards being traded or converted to steem and powered up. My best guess is that steem dollar to vests represents a large chunk of the power ups. So how much new investment is there really?

The network could conceivably balance in terms of power ups and power downs at a lower steem price. But if the downward trend continues for long enough it could shatter investor confidence and trigger selloffs. A lower price might also make people unwilling to continue contributing and curating, although given the strong steem community that may not be much of an issue.

The two biggest questions in my mind are how does the network create demand for steem and how can it generate actual revenue?

WHY POWER UP?

For people buying steem and vesting it without the intention of gaming the curation (i.e., without expecting a payout) their money is pretty much worthless in terms of vote power. The steem per mvest conversion is really high (and growing) and their share of the vests would be terribly small. So what exactly do they get out of it?

I can imagine people buying in just for fun to participate in voting but I doubt that will be anywhere near enough to support the system.

STEEM DOLLARS AS A CURRENCY

The more widely accepted steem dollars become as a currency the more demand there will be for them. Currently bitcoin is the most integrated with the existing financial system and it's also the most secure.

It took quite a while for bitcoin to achieve this, and while bitcoin has laid a lot of the groundwork for getting people to understand crypto, the steem network still has to prove its viability before more people consider accepting it as payment. And even then, as other alt coins have seen, real world integration is not something that comes easy.

Realistically I'm not sure this can be the source of the demand. Bitcoin (at least for now) already provides the utility people get from using crypto. And bitcoin is viewed as the surest bet to hold crypto long term since it's been around for almost eight years. Why switch to another currency?

STABLECOINS

One reason would be improved price stability since SMDs are supposed to be a stablecoin. If the stabilizing mechanism can work reasonably well, at least to the extent that it's significantly better than bitcoin price swings, then that might prove itself attractive enough for a faster real world integration.

While the stablecoin aspect is great, there are several significant drawbacks to steem dollars.

PRIVACY

Firstly there is a privacy problem. Bitcoin has one too -- everything is traceable. If someone can connect your public address to your real world identity then all your transactions can be tied to you.

But with steem your actual account name is in the blockchain. If you've ever posted anything it might be trivial for someone to connect your account name to your identity. Even here on the Steemit website you can click on anyone's wallet and see their holdings along with a transaction history.

With bitcoin even if you can connect a public address to someone you still have to look at the transactions and figure out what all the other public addresses are in order to understand the wallet's activities. Tracing bitcoins, as far as I know, is not exactly straightfoward either. And there are plenty of ways to enhance privacy like mixing coins or not re-using addresses.

Can similar tools be built for steem?

I don't see why not for steem and steem dollars, but you'd want to keep all your vests in one account for voting power. Prying eyes, like the taxman, will still easily be able to see how much you've made in rewards.

DOLLAR PEG

Another drawback is that SMDs are pegged to the US dollar. Many crpyto enthusiasts expect a dollar crash or large devaluation, so what would that mean for SMDs? That would hurt long term SMD holding prospects.

SMD DEMAND

If SMDs are supposed to be a general purpose currency then it's assumed that the demand for them will increase significantly. Currently there's 2.1 million of them and $2.1 million is nowhere near enough value for that sort of usage. 5.3 million steem per year get converted into SMDs so the supply depends upon the witness feed price conversion rate when rewards are paid out.

At the current feed price the market cap of SMDs will be $17.6 million after three more years. How will SMDs ever be able to support anything except a tiny market of transactions? Is Steemit planning on altering the issuance model at some point?

The SMD supply also depends upon how many people use the "convert to steem" option but this only destroys SMDs and lowers the market cap.

A large surge in demand would also drive the market price of SMDs well above $1 and the only way to combat that is for witnesses to set the interest rate to zero. This might have a small marginal effect but for a large demand relative to the small SMD market cap I'm not sure what could be done. At that point the SMD price would likely fluctuate like other crypto and SMDs would lose their stablecoin property.

SINGLE PLATFORM CURRENCY

Ultimately I think the market will choose an uncensorable crypto-currency with security, privacy, scalability, and price stability. Even if SMDs could provide all of these things why would the market want currency that is tied to one platform within one space, i.e., monetized social media?

Steem sort of has it backwards. First you need a general purpose e-money on which to build any number of platforms for a wide variety of uses. That way if one of the platforms fails or gets outcompeted it doesn't destroy the underlying currency.

Instead Steemit has constructed one social media platform and built a currency, SMDs, on top of it. If the steem network dies or gets outcompeted then SMDs become worthless.

STEEM DEMAND

Lastly, just because demand is generated for SMDs doesn't mean it will for steem. Let's say SMDs emerge as a widely used crypto-currency. Why would the steem price go up as a result? General purpose users of SMDs don't care about what steem is doing. While the steem network would likely enjoy some interest from the association there's no guarantee that this will provide the sustainability needed for payouts.

AD REVENUE

At first I thought integrating ads would be very difficult if not impossible. If it were done through the Steemit website in the form of banner ads then Steemit, Inc., would get all the revenue and it wouldn't help the network.

If it were done through advertisers buying steem to power up so they can post and up vote articles promoting their product this would mean that, for the most part, they only have to buy in once. Normally ads are paid per click. Presently there's no way to accommodate that on the steem network.

One solution I thought of however is to let people sell ad space on their articles. The links to the ads could be stored on blockchain like the article text is. Advertisers could then have their steem account deducted every time someone clicks on an ad.

Of course you'd have to make sure the clicker has a steem account (the account fee would prevent sybil attacks). And you would need to filter out multiple clicks or obvious click collusion.

But ultimately the advertisers would pay if the click rate translated into real revenue for them. And having advertisers would bring in the much needed revenue to possibly sustain the network long term.

CONCLUSION

I remain bearish on steem. I don't see SMDs ever gaining traction as a currency and I think the inflation could kill the steem price before it has less of an effect in the coming years.

The strong community however may propel the platform forward despite these concerns and new sources of revenue may be discovered.