Industrial Deal Case Study
How My Mini-Me Flags Red Flags, Stress-Tests Deals, and Writes Better LP Questions Than Most Allocators
Happy Sunday, and happy Mother’s Day to the three of you who are moms. For the other 3,859 of you (give or take), consider this your reminder to reach out to the mothers in your life. You’re welcome.
In the spirit of nurturing new critters, I thought it was only fitting to introduce my own mini-me. Mini-Leyla™ is smart (takes after her mother), fast, and crunches through pitch decks like she was born to do it. But I’m getting ahead of myself.
Today’s case study is about:
Supply chains
Nominative determinism
The former, because demand for this CRE product hinges on cross-border supply chains. The latter, because the words “conservative” and “conservatively” appear nine times in the deck, which brings us to my question of the day:
Before we dive in:
Accredited Insight is a unique newsletter: we are the only voice offering a perspective from the LP seat. We can’t guarantee you’ll avoid disasters like this one, but you’ll certainly start asking better questions after reading this publication.
By becoming a paid subscriber, you will gain access to our database of over 30 case studies and articles on everything you need to know to become a better investor. If you are a GP, this is your window into the world of capital allocators. Click the button below and chose your preferred term:
Background and US-Mexico Trade
Please remember, the deal is presented as a case study, and all relevant information has been changed to protect the identity of the sponsor. This is not investment advice.
This is an acquisition of a vacant warehouse in a border town whose economy revolves around US-Mexico trade. I’m not going to bore you to tears with another discussion on tariffs, but I will say this: cross-border trade (this applies to trade with our Northern neighbors as well) is being disrupted as we speak.
How it ultimately shakes out is anyone’s guess, but things could look very different two years from now. If you invest in logistics or industrial real estate in states where demand is driven by trade routes, it’s worth keeping an eye on the news.
The chart below (from a Barings report), shows the aggregate value of goods traded through US-Mexico Ports.
The Property
The property itself is a Class B warehouse, ~70,000 sf, with 18’ clear height ceilings, and 8,000 sf of recently renovated office space.
The Numbers
Purchase Price $8.7M (+$2.5M closing costs, tenant improvements, etc. + $2M in reserves)
Debt: 60% LTV, 24-month bridge loan, interest only (9% rate), refi in Year 2
Fees: 2% acquisition, 1% disposition, 2% of equity asset management fee. Conservative, lol. Speaking of fees:
7% hurdle (aka pref), 75/25 promote
17% IRR, 6 year hold.
The Business Plan
The business plan is to lease up the property ASAP (within the first 12 months of ownership), refinance into a permanent loan (return a portion of investors’ equity) and hold the property for several years.
Introducing Mini-Leyla™
Today, instead of giving you my take on the deal, I will show you what my AI baby can do. For a primer on what I recommend, how to get it running, and which checkpoints to use (Step 1), start here:
Where to find additional resources (reports, demographic data, etc.) to upload to your LLM:
Prompts I Used to Evaluate the Deal (in 5 Steps)
Step 2: Challenge Weak or Aggressive Assumptions
Identify and flag any assumptions that appear:
- Standardized or overly generic
- Aggressive vs. submarket or regional data
- Unsupported by the data or narrative in the deck
Highlight any red flags such as:
- Assumed cap rate compression without justification
- No sensitivity analysis on interest rate increases or operating costs
- Inflated NOI projections not supported by current rent comps or industrial demand trendsThe first red flag my little munchkin highlighted? A missing exit cap rate. And I agree - exit cap rates are the input that most often has an outsized impact on outcome. Here’s why.
The next three steps are more technical, and I strongly recommend doing them in order. LLMs think (is that what we call it?) linearly. Your results will always be more accurate if you prompt the model to work through each step explicitly and sequentially, like this: