I've gained sufficient information in last couple of weeks to write an answer myself.
As a prerequisite to the discussion, please note the difference between uniform vs non-uniform market clearing price. Non-convexities present in electricity market clearing models would lead to non-uniform prices, which would result in participants loosing money if no uplifts/side-payments are employed.
Economic interpretation of dual prices obtained from fixed-MIP
The classical economic interpretation of dual price of power balance constraint in electricity market is "the total attained increase in objective cost if the electricity demand is increased by 1 MW"; Thus, the economic interpretation of dual price obtained from fixed-MIP concern the objective cost of fixed-MIP instead of the objective cost of original MIP problem.
Hence, if the desired market clearing price is supposed to be the one corresponding to the original MIP problem [in terms of economical interpretation], then, fixed-MIP approach to obtain dual price is not suitable. In other words, the dual price obtained from fixed-MIP does not represent the cost impact of 1 MW increase of demand in original MIP problem due to the fact that the original MIP could potentially find a completely different solution to binary/integer variable in response to 1 MW increase in demand.
Additional non-convexities usually present in fixed-MIP model and their impact on dual prices
If the objective is to find a uniform market clearing price using fixed-MIP approach, one must make sure that fixed-MIP is free of all non-convexities. Most electricity market problems consist of various non-convexities in addition to already fixed binary/integer decision variables (startup/shutdown, mustrun etc.) making it impossible to find a uniform market clearing price using fixed-MIP. Such non-convexities include (non-exhaustive list):
- Transmission congestion and losses
- Minimum generation level
- Ramping constraints
- Co-optimization of energy and capacity etc.
Fixed-MIP approach in practice
In practice, various market designs employ fixed-MIP to a certain degree but a special care is always required to clarify the impact of non-convexities. Below are some examples:
- RTO/ISO type power pool markets: In this type of market, SCUC can be considered as original MIP problem and SCED can be considered equivalent to fixed-MIP. Some non-convexities such as transmission congestions and losses are handled using locational marginal pricing. Other non-convexities such as ramping, startup/shutdown etc. is handled using dedicated products which are settled via side-payments. However, even if the overall pricing mechanism is non-uniform, fixed-MIP (SCED) is used for driving the uniform part of price for energy and ancillary services products.
- Pay-as-bid to manage non-convexities: Some markets settle the energy product at a uniform price [obtained from fixed-MIP] if the offer/bid is in-the-money and use pay-as-bid if the offer/bid is out of money (mostly due to non-convexities).
- No fixed-MIP: Some other markets do not use fixed-MIP as they use alternate settlement mechanisms such as:
- Pay-as-bid for all cleared offers/bids (+ side-payments for startup/shutdown).
- Price of most expensive offer can be used if non-convexities are eliminated in the optimization algorithm by allowing sub-optimal solution.