I thought this was a great valuation problem to discuss.
Your appraisal requires the "fair market value" definition. So I'm guessing this is for estate purposes. There isn't much material difference between "Fair Market Value", the GSE's definition of value, or the IAEG's definition of value. They can all be summed up as, "Based on who the likely buyer is, what would they pay for the property if they were knowledgeable and informed, without any undue pressure?"
The "Based on who the likely buyer is..." is important in your scenario.
The fact that you've been able to determine the contract rents are "market" indicates that while the predominant ownership may be owner-user, there is sufficient rentals to come up with a market rent estimate.
I spoke to Howard about your post and he pointed out the following:
If the rents are at market, in theory, there is no leasehold value, so the fair market value of the leased fee interest should be close to the fee simple interest.
I cannot argue with this.
However, I'll add the following:
A 20-month existing lease likely narrows the pool of owner-users significantly. Is an owner-user going to purchase the house and wait 20 months before they can occupy it? Maybe, but most won't.
Therefore, in the H&BU analysis, the potential of an investor being the likely buyer has to be explored.
If an investor is the likely buyer, then the income approach takes on more significance in the reconciliation process.
Your problem may be the lack of investor-to-investor sales within your market. If you had them, you could develop a GRM based on local data.
While many appraisers pro forma the GRM (they take owner-user sales, apply a market rent to them, and use that to calculate the GRM), in my opinion, that is a last-resort technique and should be supported by actual GRMs based on investor sales.
If this assignment were mine, I'd go to competing markets where I could find investor-to-investor sales and use the rents to calculate a GRM. After doing a number of these, I'll come up with a range (let's just say the range is 150 to 180). Then, you could pro forma owner-user sales with market rents to see what their indicated GRMs are. My "guess" is that it will be higher (if it were lower, investors, in theory, would start buying these homes and renting them out). Let's just say the pro forma range is 170 to 200.
You can see what we have here under this example: an intersection of the two data sets between 170 and 180. It would not be unreasonable to pick a GRM for the subject closer to the higher range, nor (IMO) unreasonable to go a bit (185?) higher.
Using the GRM, you can complete the income approach. Again, based on the data in the example, the sales comparison approach will likely conclude a value that is higher than the income approach (purists will argue that a GRM is not a true income-approach; they'd be correct but for residential work, it is a recognized technique).
If I believed the likely buyer of the property is an investor (given the lease terms), then I'd give more consideration to the income approach vs. the sales comparison approach. My final opinion of fair market value may be below the indicated range of the sales comparison approach.
Alternatively, if I believed that, despite the lease, an owner-user would compete for the property, I might give both approaches equal consideration (you get the picture).
One last thing for you to confirm (and Howard pointed this out to me): Is there a cancellation clause in the lease that is triggered upon the sale of the property? If so, and if this is an owner-user market for the subject, I'd still do the income approach but reconcile to the sales comparison approach.
I won't get into the technicalities of being sure to match the risk-profile of your subject with the comparable rentals you use to extract the GRMs; if this is for a simple estate assignment, then while you wouldn't want to compare a triplex GRMs to use for your SFR, I think you have flexibility in comparing SFRs with some differences in their risk profile to use for this assignment.
Good luck!