Staking Rewards is proud to present the most comprehensive staking data study ever made for the Tezos Blockchain. The research report is conducted by Blockwatch Data, who are leveraging their comprehensive data sets to explore delegator lifecycles and behavior, the impact of exchanges, and questions like:
- Are bakers and delegators able to build trustful long-term relationships?
- Do Delegators accumulate for compounding rewards or are they selling of stakes?
- What challenges for Tezos does the data suggest?
Tezos is a proof-of-stake blockchain with a two-year history of active participation. Today, in July 2020, there are around 440 active validators and 56k delegators on the Tezos network. We were curious to see how staking on Tezos evolved over time and especially whether the recent introduction of custodial staking on large exchanges like Coinbase has contributed to significant shifts in delegation behavior.
The Underlying Data
For this study, we used data from our TzStats Data API which drives our Tezos Block Explorer TzStats. We look at all historic baker and delegator snapshots taken by the blockchain every 256 blocks (approx. 4.25 hours) to determine baking rights. With 63+ million rows, this is the most comprehensive and detailed dataset on Tezos staking activity in existence. The dataset contains all active bakers with rolls and all non-zero delegators at each snapshot. We further pre-processed the raw data and labeled all delegations to (1) identify sold-out delegators whose balance dropped below 1 tez and (2) mark delegators belonging to the Tezos Foundation or known custodians in order to exclude both groups from our analysis. Out of the current 56k active delegators, only 42k have a balance larger than 1 tez.
Staking in Tezos
Before we start, let’s recap how Tezos staking works. Tezos uses a distinctive form of staking called Liquid Proof-of-Stake, which separates participants into two groups — block validators (bakers) and delegators. While bakers have to lock-in (freeze) deposits to encourage honest behavior, delegators can freely spend and redelegate their coins. Thus, coins always stay under a delegator’s control. What’s delegated is merely the right that is attached to coins. It defines the staking power of a baker (rolls).
The more rolls a baker owns relative to network-wide rolls, the more blocks they can bake or validate (endorse) for rewards. In addition to rewards, bakers can earn transaction fees from blocks, but these are too low to matter right now. Total rewards are capped at 80 XTZ per block or 42M XTZ annually. This currently translates to ~5% inflation. Since rewards are fixed, the inflation rate will slowly decrease over time unless a large amount of tez gets burned. Inherent to staking is that non-stakers are diluted by inflation. On the other side, if not all coins are staked, bakers and their delegators can make a profit equal to the dilution of the non-stakers. Right now, the staking rate in Tezos is at 80.32% (666M XTZ), which yields an estimated gain of around 1% for stakers.
If not all bakers perform their job on time, the network becomes unhealthy and less rewards are paid. To put this in numbers, 4.7% (1.96M XTZ) of past year’s possible rewards have not been paid due to unreliable bakers. Since the Carthage network upgrade in March 2020, rewards have become more dynamic. Although this makes attacks less attractive and encourages better baker performance it has an additional dampening effect on inflation which will likely be more pronounced this year.
A common misconception about Tezos is that all staked coins are out of circulation, i.e. they cannot move or be sold next block. In fact, only baker deposits, fees, and rewards for the past 14 days are locked (frozen). The total amount of frozen capital across all bakers is at most 6.2% (52.4M XTZ) in deposits and 0.2% (1.6M XTZ) in rewards.
Together, all 442 currently active Tezos bakers manage 16.2% (136M XTZ) of their own capital and receive 63.8% (537M XTZ) in delegations. 5.9% (49M XTZ) is not activated yet, 1% (9.2M XTZ) uselessly stakes with inactive bakers, and 20% (169M XTZ) does not stake. 421 of the 441 active bakers hold enough stake to receive at least one roll, which lets them participate in the lottery for blocks and endorsements.
Tezos Network Totals on July 6, 2020
|Active Staking Capital||672,992,992 (79.95% of total)|
|Activated Fundraiser Capital||562,275,255 (66.8% of total and 92% of fundraiser)|
|Unclaimed Fundraiser Capital||49,179,622 (5.9% of total and 8% of fundraiser)|
|Active Baker Capital||136,256,254 (16.2% of total)|
|Active Delegator Capital||536,736,738 (63.8% of total)|
|Frozen Capital||63,968,000 (7.6% of total)|
|Max Annual Inflation||42,048,000 (5.0% vs current total)|
|Current Inflation||40,091,070 (4.8% vs current total)|
|Bakers with Rolls||421|
|Inactive Delegator Capital||4,782,064|
|Inactive Baker Capital||4,393,030|
To have a greater chance to earn rewards, bakers compete for rolls. This puts delegations into the spotlight. Bakers have to leverage their capital to successfully attract and bind new delegators and receive more rolls. Long-lasting baker-delegator relations thus play a key role as a competitive advantage for bakers. Let’s explore these relations in depth.
Tezos Baker-Delegator Relationship Insights
Delegator growth in Tezos has been reaching new peaks almost every month. Currently, around 80% of the entire Tezos supply and 56k accounts (42k with more than 1 tez balance) participate in staking. This represents a growth of about 300% since May 2019 (~13.7k) and is the highest staking ratio in all major PoS networks at the time of posting. Out of 441 registered bakers around 120 constitute public bakers, i.e. delegation services who compete for delegations.
Staking creates a powerful social bond between bakers and delegators that goes far beyond the technical aspects of a blockchain system. Bakers have to first earn the attention and then continuously build a trust relation with delegators. Delegators, on the other hand, expect to receive staking rewards on a regular basis but may switch to a different baker at any point in time for any reason. Intuitively this relation reminds of that between couples, so we look at staking through this lens. Are bakers and delegators able to build trustful long-term relationships? How often and when do breakups happen and why do delegators change their minds?
We start by analyzing the overall influx of new delegators into Tezos to shed light on how staking evolved. Figure 1 displays the short 2 years of existence in which Tezos went through three generations of growth. We call them the Early Backers (Jun’18 – Mar’19), the 2019 Boomers (Apr’19 – Oct’19), and the Brrrr Generation (Nov’19 – now). Each generation started with an up-tick in delegator growth and a parallel price increase before both peaked and ebbed out. Towards the end of a generation, Tezos had gained 2-3x more delegators. The most recent growth cycle is not over yet, and we are currently witnessing the largest delegator growth so far, with an all-time high in May 2020 (+7.5k). This is 10x the growth we saw at the low last September.
Next, in Figure 2, we look at the rate of breakups. We use “churn rate”, a common metric that tracks how many customers stop using a product over a specific time period. In our case churn expresses the rate at which delegators leave their current baker each month. Churn also implicitly tells us how long delegators remain loyal, in other words, how quick bakers need to find new delegators to stay in business. Delegator churn in Tezos is surprisingly low at only 2 to 10 percent, suggesting delegators stay between 10 to 50 months on average. Over time, delegator churn fluctuates in waves which are aligned with growth displayed above.
As more and more bakers emerged during the first two months after the network’s initial launch, many early delegators experimented with the protocol, but rather quickly settled with their baker of choice. We observe that each time delegator growth rises, churn rises as well, but the intensity of churn seems unrelated. No matter how many new delegators arrive, more than 90% are loyal. Who are these loyal delegators?
Figure 3 shows the percentage of loyal delegators by the month of initial delegation, i.e. those that stick with their original baker until today. Each bar represents one cohort of people who started delegating in the respective month. Unlike growth and churn, loyalty displays surprisingly little variation between cohorts, but especially between delegator generations where we don’t find a clear trend. Around 40% of all long-term delegators remain loyal to their choice while 60% either switched baker or sold their stake. It almost seems as if 40% is a magical constant in Tezos. Note that the youngest generation had the least time to change their minds, hence the peak in June 2020 is most likely biased.
When about half of all delegators break up at some point, how long do individual relationships last? Are there any differences between generations? Figure 4 shows the average time to divorce for each cohort both in absolute terms (number of days on the left-hand axis) and relative to the maximum potential time if the relation had lasted until today (in percent on the right-hand axis). From all delegators who broke up, early backers experienced the longest-lasting relations with over 150 days on average. Early backers are clearly the most patient. In younger generations, this number has been steadily decreasing. Today, break-ups happen much more rapidly and increasingly even before first rewards are paid out. Is it really that young delegators can’t wait? Bakers have repeatedly noticed this trend before, but has it existed historically, and is the trend accelerating?
Figure 5 shows that a certain degree of impatience has always existed among delegators (10-20% of delegators always used to switch before payout), but in the most recent generation, early break-ups increased significantly. In May 2020 the trend hit an all-time high at a startling 37%. This should be concerning to the Tezos community as it clearly shows that more recent delegators do not fully understand the intricacies of the protocol or simply don’t care. We encourage the community to take this trend seriously and improve the user interface design of wallets and block explorers as we did with the recent TzStats v2 release that displays a new delegation timeline now.
One last question remains: What do loyal delegators do with their rewards? To answer this question we split delegators into two groups — accumulators with a steadily increasing balance who never spend their original stake nor their rewards, and opportunists who from time to time sell off some stake. This time we look at all currently active delegators, but we exclude those accounts owned by known custodians or the Tezos Foundation bakers. Out of all active delegators with a balance larger than 1 tez, the vast majority of 31k (72%) are accumulators and only 12k (28%) are opportunists. Over time we observe that the older a delegation, the more likely it is that the owner sold some stake. Figure 6 displays the accumulator/opportunist share for each cohort.
Although the absolute count of new delegations has increased sharply (see Figure 1) the trend to turn into an opportunist is only increasing linearly with time. Well above 28% of all delegators in each cohort have been accumulating rewards even though many of them started their delegation more than a year ago and Tezos’ price has increased 2–10x since delegation started. This demonstrates solid confidence among Tezos’ stakeholders across all generations.
An increasing share of Tezos delegators is showing considerable confidence in their respective bakers and the network overall. As staking activity grows, the number of less knowledgeable delegators is rising which leads to increased early break-ups and higher than necessary churn. We believe improved user interface design and more education can help younger delegator generations better understand Tezos staking in general and payout delays in particular. What’s more important is that we as a community have to embrace the fact that Tezos forms social relations on top of a technical system. Like in any good relationship, communication is key. That’s why we at Blockwatch work on new tools to improve the ways bakers and delegators interact to foster more trust and loyalty.
About the Authors
Blockwatch Data Inc. is an advanced blockchain analytics provider dedicated to helping you understand what’s happening on blockchains. For Tezos, Blockwatch provides the popular Tezos block explorer TzStats, the next-generation Tezos indexer TzIndex, and a powerful on-chain data API.
Connect with Blockwatch Data on Twitter @blockwatch_data.