As a follow-on to Part 1 of the Economics behind RWA, if we view the cost of capital from the lens of the borrower (especially when the cost of capital from Maker is zero), the interest paid can be viewed as a tax. Thus a similar economic theory surrounding taxation that can be found in the Laffer curve, will also be seen in the portfolio of collateral that is on-boarded within the Maker ecosystem. As such we can find and use a similar metric to help secure what is appropriate pricing from both an individual collateral perspective as well as the portfolio as a whole.
If we set the "tax" too low, MKR gets basically nothing and takes all of the risk. If we "tax" too much, MKR has zero risk and no collateral. Point being and subject to the elasticity of demand for each collateral type, if we increase the cost of capital, borrowers at some point will identify the costs as too economically expensive and will go somewhere else. In the crypto world, some of that can be "captive" to Maker; however, in a mature capital market, real world assets already have efficient capital sources. MKR Governance must be aware that pricing for Real World Assets must always remain competitive with the traditional financial system, less off-chain lenders will pay down their debt and move the loans to the traditional analog banking sector.
The critical take-away here is that for DAI to truly be a synthetic US Dollar, it must incorporate a portfolio of assets that must ultimately be as uncorrelated as possible for maximum stability. When a proper balance is found between off-chain and on-chain assets, the DSR should then be used to spur or contract DAI demand. At this point, the "tax" will be optimized for on-chain and off-chain assets. These offsetting forces causes a massive win for MKR.
Summary: Costs of capital for Real World Assets when brought on as collateral are price (interest rate) sensitive comparative to the existing banking sector.
I am presently advancing a real world assets structure for consideration by the MKR community.