Why a lack of advanced modeling techniques? Anyone who uses probabilistic modeling justification for this?
From what I have seen, real estate modeling (at least on an asset level) seems to be relatively rudimentary in comparison to almost all other financial sectors, could anyone possibly speak to why very few firms utilize advanced modeling techniques and probabilistic modeling in real estate? Or if your firm does use these techniques, what programs are you using and in what context are you using this?
Because there are so many variables in a real estate deal that are subject to change, that running probability analysis above and beyond pretty simple sensitivity matrices becomes a dumb mental jerk-off exercise that is not grounded in reality.
Take a traditional development deal. Things that are almost guaranteed to change from start to finish--construction duration, actual start date, final hard costs, the size and rent on the initial pre-lease, the pace and lumpiness of the cash flows from the balance of the lease-up, the interest carry on floating debt, the property tax bill at stabilization, the actual value of concessions, how much density you actually get approved...there are a billion items that change. Better to come up with a concept you think stands a reasonably likely chance of being built and run that on general market assumptions, than do a circle jerk.
Some shops do it though, I think it is an incredible waste of time and analysis by paralysis.
Would you be able to provide examples of some of the shops that are doing this, and if you know which variables they are looking at in this context?
Complete waste of time, totally agree. You really only need to have one case in real estate investing in my mind, the downside case. If you are comfortable with this case and still think the deal is good, then you don't need to spend time doing a ton of what-ifs on every other slight change in variable.
My shop ran insane sensitivity analyses for years. We finally stopped doing it because the outcome was always the same. Either 1.) If everything goes wrong we are fucked 2.) If everything goes right we will make a killing. In reality we always fell somewhere in the middle.
best comment I've read in a while.
I think there is an important distinction to make here: portfolio level vs asset level. In regards to a asset-level analysis, these types of sensitivity's or analyses are largely a waste of time, however, on a portfolio level. I work on the asset level side but I would venture to assume this type of analysis is being done on a portfolio level at some shops. Can anyone speak to this?
Even at fairly large funds, this doesn't happen as often as you think (we don't do this deep of an analysis). Again, you're basically trying to predict a range of values, so rather than wasting time doing a billion+ combinations of different variables going up and down, most shops pick one metric and piggyback off that. For example, most groups will play with the exit cap rate and run down, base, and up side cases based on that. It gets you to the same place as if you do a ton of different rent growth projections, starting market rents, cap ex contingencies, etc. There are a lot more variables that can go into a real estate deal than almost any other asset class, so it's a lot more head trauma to do some of these 'advanced' metrics. If someone did this and was pitching us as the LP, candidly I would think they don't know what they're doing. I would have a much higher confidence level in a developer that walks into the meeting with a simple range of values based on one or two variables changing and then being able to speak to the process, as opposed to giving me a simulation that's taken into account all 1,956,345,754,342 possible combinations that are all based on pure speculation anyways.
Real estate is, by nature, a much more opaque and inefficient market than others. At the end of the day your underwriting is always going to rely mostly on assumptions.
I could see the value of doing a monte carlo analysis to see the probabilities of different IRRs, values, or DSCRs. Does anyone do this?
Unless a particular set of outcomes are fundamentally skewed, I don't see why it would be useful.
If you want a practical guide for doing this in the RE industry, I highly recommend https://www.amazon.com/Flexibility-Estate-Valuation-under-Uncertainty/d…</a">this book.
That said, I've talked about this with people who work at major REPE and development firms that invest many billions of dollars, and they don't seem to do anything more complex than scenario analysis. The authors of the linked book are academics who are trying to increase the sophistication of the industry.
Academics who probably have never developed/managed/investing in a single development project. The most successful real estate investors/developers I know do simple back of the napkin analysis to determine project feasibility. They generate tremendous alpha through their ability to buy well located land below market and/or cut leases at above market rates with favorable legal terms to landlord.
It depends on asset class, if you're doing an office value-add with a lot of tenants and roll it could get pretty complicated with renewal probabilities. While buying NNN lease properties their isn't much to do. Some large Gateway players also have complicated models. If you doing a dev deal in New York and you got all your closing cost and tax assessment assumptions it can get very complicated.
Yup, impact fees, permitting fees, taxes, sitework, and MEP are always a bitch to model. I never see any cost savings here...
I cannot speak on behalf of other financial sectors, however, rudimentary modeling is sufficient because underwriting is primarily done on general assumptions. At the previous REIT I worked for and my current job within a Real Estate Advisory practice, we use ARGUS for Cash Flow modeling and paste into Excel models for the analysis of an Acquisition, Development and Disposition. At the previous REIT we maximum ran 4 scenarios and would do modeling based on 10 years (hold period). For example, Tenant renews at the expiration of a 5 year lease, Tenant vacates at the expiration of a 5 year lease and takes 6, 12, 18 months for lease-up.
Exit cap-rate is typically the main variable, which is impossible to predict. Some people peg exit cap-rates to a risk spread over forward treasury curves, but even this is farting in the wind.
Definitely is, baked into all of our models we would always show an estimated exit cap and a 25 basis point spread each way.
Do you think a 25 basis point spread in both directions is really enough? What type of holding period would you be looking at where you'd feel comfortable with such a small change in cap rates?
I'm currently in a model from one of the largest developers in the country, based on the created date in the info tab, its 20th b-day is coming up at the end of the month.
On that note could it be argued that the real estate industry is largely stuck in the past terms of complexity of analysis or would you also agree that this would simply lead to analysis paralysis? I understand that it's difficult to quantify all the variables but having the ability to layer 30+ variables for example each with their own respective std. dev. and range of possibility to come to probabilities of success or failure would seem hugely beneficial especially when pitching to investors or investment committees possibly
I wouldn't think it would lead to analysis paralysis if you have the model built flexibly enough that you are primarily entering inputs and getting the outputs with a range of valuations and probabilities.
If you think it is 30 variables you are laughably mistaken. There is more than 30 variables in site inspections alone in a small development project. The average RE project I have has upwards of 150+ variables from project inception to completion, and these aren't even projects on the level of what is being discussed here. It isn't a reality that the industry wouldn't use more sophisticated analysis tools, it is just that the time it would take to do so is insane.
Voluptatibus odit accusantium iste illum qui quae dolorem sequi. Ipsam error dolorem tenetur dolorum consequatur quia. Similique aliquam ad quis quasi ut sed aperiam. Nesciunt eveniet saepe qui id.
Et eum et maxime doloremque et est harum. Officia aut iure hic reiciendis. Id pariatur vel aliquid quasi quas laboriosam mollitia. Ab autem voluptates voluptatum et.
Enim aspernatur eius voluptate sequi nihil saepe. Sit omnis atque maiores ipsam. Et reprehenderit maxime dicta suscipit cum.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...
Sed itaque ut quidem iusto odio minus. Iste aut dolorem et iure ut sed exercitationem a. Tenetur aut dolorem id est quae magni beatae. Doloribus velit qui sequi incidunt quo voluptatem.
Dignissimos sed ut autem quibusdam cum. Cupiditate dolores voluptatem accusamus dolores. Voluptas consequatur qui voluptas voluptatem earum nobis. Quis quia saepe ab quia ea. Explicabo est asperiores sit quo ullam. Repellat ullam molestiae accusamus error.
Eligendi rerum quis non sed quae. Qui ut reprehenderit laboriosam doloribus. Omnis dolor est hic aliquam ullam omnis nobis qui. Repellendus explicabo officia et delectus illo.
Enim aspernatur eveniet ut omnis unde sit occaecati. Nemo rem debitis vero ut minima qui. Incidunt architecto consectetur et et nostrum rerum. Qui corrupti est quia consequuntur. Perferendis repellat nisi corrupti suscipit odit.