TLDR; We experimented with the stablecoin pool swap fees to understand which fee is ideal.
Recently, we have been experimenting with changes to the fee rate on the stablecoin pools. We set the USDT/BNT, USDC/BNT and DAI/BNT pool fees to 1%, 0.5%, and 0.2%, respectively, from 0.2%. The results obtained indicated the the ideal swap fees for the pool were 1.0% for the USDC and USDT pools and 1.5% for the DAI pool.
The plots show the average daily volume and fees accrued at different swap fees for all pools.
Very interesting over all, why is 1% more ideal than .2% for USDC? are we looking at Volume to Fees Ratio here exclusively ? at 0.2% we are seeing close to the same revenue with higher volume which might translate into swaps elsewhere in our pools.
tenzent makes a really good point here. If the “fee return curve” has two maxima at different %s, with (presumably) the lower one supporting a higher volume, it would be better to choose the lower one to support other trades in the bancor network.
@NIX is probably the best to expand on this, but I should note that the graphs above don’t include market share above and below the swap fee changes and capture relatively long periods of time. The fee was set at 0.2% on the stablepools during periods of higher volume/volatility. Therefore, although a good benchmark to analyse pool performance, average daily volume and fee at different swap fees shouldn’t be taken solely into consideration when deciding what fees to set the pools.
Furthermore, the experiment is ongoing and the goal to maximise profit for all LPs in the USDT, USDC and DAI pools is still being pursued, with more data analysis to come in a near future.