Congratulations Marc on producing this detailed analysis, and thanks for posting it here. I would like to comment on a number of points.
Familiarity bias.
You consider familiarity bias so important that you address it as your first point, yet I fail to understand what you are saying about it. The apartment I can see out my window, doesn't appear on the radar of an investor in New York. To me this says that the market in that apartment is likely to be less efficient than the market in a widely traded share, and so there is a greater chance that it is mispriced. Possibly underpriced, possibly overpriced. To me this gives more opportunity to well informed property investor to find a bargain than would be found in a widely traded share.
Liquidity
You say that property is not liquid. That is true. You say that therefore an investor should seek an illiquidity premium. I disagree. If an investor has funds that are not suitable to be tied up long term in an illiquid investment then they should not be invested in property. If an investors funds can confidently be tied up for the long term then because short term investors are out, the investor may expect to gain a better return. This is not the same as a liquidity premium, anyone who needs their return in 5 years, but invests in property because he gets an “illiquidity premium” is a fool. Anyone who is confident he can tie his money up for 25 years should invest in the asset with the best expected return over that period.
Starting Valuations
You say “Starting Valuations clearly matter when forming a view about expected future returns in both equity and real estate investments” Absolutely. Although I believe that this is more relevant for property investments, because the market is shallower.
Residential properties vs Rental properties.
On page 24 you compare the rental yield on properties in Dublin 6 with those in Ballyfermot. I would suggest that including properties in D6 or other similar areas distorts the analysis of the returns to property investing. While every share is bought with a view to profit, not every property is. The price of property in D6 is not driven by its investment potential, its driven by social chachet, which has no place in investing. Including such properties in an analysis of property as an investment makes the returns look very much worse than the actual returns experienced by those who buy property as an investment rather than as a home. Of course I realise that it would be difficult to exclude these type of properties.
This is my most important disagreement with your analysis. I suggest that an analysis of an investment case based on properties many (the large majority ?) of which were not bought for investment purposes, is futile.
Bias
There are a number of hints that this research is seeking to come to a predetermined conclusion rather that being open to finding among a number of possible conclusions.
The quotation from Schiller at the beginning for a start.
In your conclusion Section 3.4 page 25 you say that the expected real return for Irish residential property nationally is around 2%,. It is only in 3.4.1 that we see that this refers to capital appreciation only. I don't think that you add this anywhere to the rental income to obtain a total return. A curious omission.
Your assumed Rental Income for Irish Residential property is 3% to 4.5% Gross. Page 25. You really need to get out more Marc. That is even contradicted by the details on page 24 where you tell us that the yield in D6 is the lowest in the country at 3.6%. I suggest that yields in excess of 8% gross are widely available in the market.
As a by the way calculating an expected return based on historical performance is not something I see as a valid exercise, you give some reasons for this yourself on the bottom of page 22.
Two small points
It seems that in your graphs you compare equity returns net of dividend with property price returns. I suggest that this is less than compelling. Obviously it is total returns that matter in both cases. The split between capital and yield in each asset category is not necessarily comparable.
Page 22. I think the labelling of the X axis is a typo.
Thanks again for sharing your work, comparing my opinions with your work has helped me clarify me ideas about property investing.