Tax tokens refer to tokens that users will not receive in full when trading on a decentralized exchange (DEX), and some tokens will be destroyed or distributed as dividends. For example, if the price of UNIBOT (5% tax on transactions) is 100 USDT, when a user spends 100 USDT to buy it, he can only receive 0.95 tokens, and similarly, when selling one, he can only receive 95 USDT.

Centralized exchanges (CEX) do not have the function and mechanism to collect taxes. For non-taxable tokens, if there is a price difference, there will naturally be arbitrageurs to level the difference. However, for taxable tokens, in theory, the price difference will be in a state of superposition of positive premium and negative premium. Let's take UNIBOT at 100 USDT as an example:

For sellers, negative premiums on exchanges are acceptable as long as they are within -5%. For example, at a -4% premium, they can still sell for 96 USDT on CEX, which is still a profit compared to the 95 USDT on the chain.

For buyers, they can accept a positive premium of less than 5%, that is, CEX transaction prices below 105USDT are all cost-effective compared to the on-chain prices.

Therefore, the positive or negative premium of CEX may be the result of the game between the above two forces, and eventually reach a balance. Based on this, Odaily Planet Daily makes a hypothesis and will verify it below:

If traders are bearish on the future market and the selling power is stronger, the exchange will show a continuous negative premium state and the coin price will fall;

If traders are bullish on the future market and the buying power is stronger, the exchange will show a continuous positive premium state and the coin price will rise;

UNIBOT Data Analysis

basic rules

Analysis target: Taking UNIBOT as the first analysis object, the token transaction is subject to a 5% tax, which is distributed to LP, coin holders and the team respectively;

Time dimension: 1 hour is used as the analysis dimension;

Time span: nearly 1,600 hours, from about 8/23 to 10/27;

Price value: DEX takes the average of all transaction prices within an hour (from Dune's dex.prices); CEX takes Bitget data and takes the average of the opening, closing, highest and lowest prices within an hour;

Premium calculation: Premium rate = (CEX-DEX)/DEX.

Calculation results

Among the 1,600 hours, the premium exceeded +5% or -5% in 790 hours, accounting for nearly 50%. This indicates that the market power is too strong (overbought or oversold), or the market power is too weak and dominated by exchanges or market makers.

In order to explore the relationship between positive and negative premiums and trends, Odaily Planet Daily records premiums above +3% as positive premiums (orange in the figure below) and premiums below -3% as negative premiums (purple in the figure below).

In a market trend, there will be a continuous unilateral premium without the existence of a premium on the other side. The interval with mainly positive premium will be marked with a blue background, and the interval with mainly negative premium will be marked with a yellow background.

As can be seen from the figure, there is an obvious relationship between the positive and negative premium ranges and the price trend, and we can get conclusion ①, which is completely opposite to the law speculated at the beginning of this article:

① When there is a positive premium, prices tend to fall; when there is a negative premium, prices tend to rise.

In addition, there are several other derived conclusions from the figure. The premium range selected in this article is relatively large, and there are small reverse premium ranges in the large premium range. For example, in the 9/17~9/22 range, there were several small negative premium ranges in the range dominated by positive premiums. The price trend rebounded in this range, but the overall trend did not reverse.

The long-term premium changed between 9/22 and 9/27, and the trend reversed. Then on 10/2, the premium changed again and the trend reversed at the same time.

Therefore, the conclusions are as follows:

②The trend is related to the long-term premium situation. Short-term premium means reverse development, but does not change the overall direction.

③After the long-term premium conversion, the overall trend reversed.

So, is positive or negative premium a sufficient or necessary condition for price movement? The answer is neither:

The overall trend from 8/23 to 8/28 was upward, with an increase of more than 30%, but there was no positive or negative premium range, which negated the necessary conditions;

There were small premium ranges around 10/4 and 10/24, but the price went sideways, negating the sufficient condition;

Therefore, we have conclusion ④ that positive and negative premiums are strongly correlated with trends, but this is neither a sufficient nor a necessary condition.

Further data verification

AIMBOT

The AIMBOT token of the automatic dog-beating robot Aimbot is used as the research object. The token also charges a 5% tax on transactions, but has another feature: a 5% tax is also charged on transfers. This feature prevents arbitrageurs from moving money, and will also curb the on-chain and off-chain transfers of traders. Transactions on and off the chain will become "stand-alone".

As can be seen from the above figure, AIMBOT no longer has obvious positive and negative premium ranges. The "stand-alone" state makes the positive and negative premiums very random, but when the range is formed, the price trend is still in line with the above. The new conclusions are as follows:

⑤On-chain and off-chain transfer channels are a necessary condition for CEX to maintain a premium.

BANANA

The token BANANA of Telegram transaction BotBananaGun is used as the research object. The transaction of this token is taxed at 4%. The limit of positive and negative premium is modified to ±2%. Other conditions are consistent with UNIBOT.

It can be seen that the correlation between the positive and negative premium ranges and price trends still exists, but the price fluctuations are not particularly obvious, which may be related to the fact that BANANA is a popular new currency.

Compared with AIMBOT, BANANA does not charge transfer fees, so the premium range has become wider and more significant again, confirming the conclusion ⑤.

in conclusion

In summary, there is an obvious correlation between the long-term premium range and the price trend. The positive premium falls and the negative premium rises. The reversal of the premium is also one of the signals of a trend reversal, but it still needs to be identified and judged in combination with the characteristics of the token itself and the situation of the exchange.