An investment committee memo that cannot be wrong is not worth reading. Most memos are summaries: here is the property, here are the numbers, here is the recommendation. Every sentence in them is true and none of it is testable. A memo earns its place when it states a thesis specific enough to be proven false, names the two or three assumptions the entire return rests on, and says in advance what evidence would kill the deal. That structure changes what the committee argues about, and it leaves a record the firm can grade later.
Key Takeaways
A summary cannot be wrong, which is the problem with it. A memo should make a claim that reality is capable of contradicting.
Every deal's return rests on two or three load-bearing assumptions. A memo that lists twenty risks and ranks none of them has said nothing.
Kill criteria belong in the memo before diligence begins, not in the post mortem after the asset underperforms.
Precision is the price of accountability. "Market rents support the plan" cannot be graded. "$9.00 NNN by month 36" can.
An IC memo is a forecasting record. A firm that never writes down what it believed can never learn whether its process works.
What should an investment committee memo contain?
A real estate investment committee memo needs seven parts: the thesis, the basis and pricing, the two or three assumptions the return depends on, the evidence behind each one, the downside case, the kill criteria, and the recommendation with a stated level of conviction. Description of the property is context, not content.
Most IC memo structure guidance stops at the section list, which is why so many memos are long and empty. The list is not the hard part. Every section has to make a claim that could turn out to be false. A property description cannot be false. A statement that 40,000 square feet releases at a stated rent by a stated month can be.
Section | The claim it must make | How it can be falsified |
|---|---|---|
Thesis | Why this asset produces this return, in two sentences | The mechanism does not occur |
Basis and pricing | What we pay per unit of income against verified trades | Comparables print below our basis |
Load-bearing assumptions | The two or three inputs that carry the return | Sensitivity promotes a fourth input |
Evidence | The leases, comps, and quotes behind each assumption | The source documents say otherwise |
Downside case | What the deal returns if the main assumption misses | The outcome is worse than shown |
Kill criteria | The findings that end the deal, set before diligence | A criterion is hit and the deal proceeds |
Recommendation | The decision, the conviction level, and who owns it | The vote and the memo disagree |
Why is a summary memo weaker than an argument?
A summary memo is weaker because it cannot be checked. Karl Popper's standard for a scientific claim was that it must forbid something: a statement compatible with every outcome carries no information. Most deal memos in commercial real estate are compatible with every outcome, which is why committees approve them and learn nothing.
Watch the room. A summary produces a discussion about the summary: the roof, the submarket, the sponsor, in no particular order, because nothing in the document says which question matters most. An argument sends the hour to the variable that decides the deal.
The pairs below are illustrative.
Summary language | Falsifiable language |
|---|---|
The submarket has strong fundamentals | Vacancy is under 6 percent and has fallen four straight quarters; the plan fails above 9 percent by month 24 |
Rents are below market | In-place rent is $6.50 NNN; four verified leases signed between $8.75 and $9.25; we underwrite $9.00 by month 36 |
The exit assumes modest cap expansion | Exit cap 6.50 percent against a 6.00 percent going-in; at 7.00 percent the deal returns capital and nothing more |
Risks include lease-up timing | If two leases are not signed by month 18, we hold to stabilization rather than sell |
The left column is safe: nobody is ever wrong for having written it. The right column exposes the author, and that exposure is the value of the document.
How do you find the assumptions a deal's return depends on?
You find them by running each input to failure and watching which ones move the outcome. Two or three will dominate and the rest will not matter. In most stabilized acquisitions the dominant inputs are the exit cap rate, the achievable rent on rolling space, and the timing of that roll. Rank them and write the ranking down.
Every figure below derives from the stated inputs.
A 100,000 square foot industrial building at $10,000,000, or $100 per square foot. In-place rent is $6.50 NNN, or $650,000, less $50,000 of non-reimbursable expense and credit loss, for a going-in NOI of $600,000 and a 6.00 percent cap rate. The stack is $6,000,000 of debt and $4,000,000 of equity. The plan: 40,000 square feet expires in year three and marks from $6.50 to $9.00, adding $2.50 on 40,000 square feet, or $100,000 of NOI. Year five NOI is $700,000, exited at 6.00 percent.
Scenario | Year 5 NOI | Exit cap | Exit value | Value over $10.0M basis |
|---|---|---|---|---|
Underwritten: $9.00 rent, 6.00 percent exit | $700,000 | 6.00% | $11,666,667 | $1,666,667 |
Rent lands at $8.00 | $660,000 | 6.00% | $11,000,000 | $1,000,000 |
Rent holds, exit cap widens 50 bp | $700,000 | 6.50% | $10,769,231 | $769,231 |
Both miss | $660,000 | 6.50% | $10,153,846 | $153,846 |
Read the third row. A 50 basis point move in the exit cap, which no one controls and no one can verify at closing, removes $897,436 of value, or 22 percent of the equity check. In the fourth, a one dollar rent miss plus that same cap move removes 91 percent of the value created over basis. The deal survives on paper and returns close to nothing after fees.
So the memo does not open by describing loading docks. It opens like this: this deal is a bet that 40,000 square feet releases at $9.00 NNN by month 36 and that the exit cap is no wider than 6.50 percent. If one fails, the deal returns capital. If both fail, it loses money after fees. Everything else here is context for those two numbers.
That is a claim a committee can attack. The mechanics run through the pro forma assumptions buyers should challenge and sensitivity analysis versus a guess.
What are kill criteria and when should they be written?
Kill criteria are the findings that end a deal, written into the memo before diligence begins and stated as thresholds rather than concerns. They work because they are set while the author is still neutral. Written after the team has spent weeks and real money on third-party reports, the same findings become problems to solve rather than reasons to stop.
Gary Klein's premortem, published in Harvard Business Review in September 2007, is the cleanest version of the technique: assume the deal has already failed, then generate the reasons. That exercise produces criteria instead of anxieties. Continuing the example: if the two best verifiable comparable leases signed below $8.25 NNN, we retrade or withdraw. If tenant improvement and commission quotes exceed $25 per square foot on the rolling space, net effective rent does not support the $9.00 mark. If the lender sizes below $6,000,000, the equity check grows and the case above no longer clears our threshold.
A memo that calls leasing risk a consideration gives the committee nothing to enforce. A memo that names $8.25 gives it a trigger.
How does an IC memo become a record the firm can learn from?
It becomes a record when its assumptions are specific enough to score at exit. Philip Tetlock's work on forecasting, summarized in Superforecasting, holds that accuracy improves only where forecasts are precise enough to be graded and where someone keeps score. Vague forecasts feel safer and teach nothing, because they can never be marked wrong.
Real estate has slow feedback. A five-year hold puts five years between the belief and the result, and by then the author has been promoted, has left, or remembers the deal differently than the document does. The memo is the only unedited copy of what the firm believed when it committed capital.
The practice is unglamorous. At exit, pull the original memo, set the underwritten assumption beside the realized outcome, and log the gap. Do that for twenty deals and a pattern appears that no single deal reveals: rent marks run optimistic, lease-up runs long, or exit caps are underwritten to the day of purchase rather than to a cycle. Only firms whose memos said something specific enough to score ever get that calibration.
A memo that cannot be wrong cannot be checked, and a firm that never checks its memos is not underwriting. It is decorating.
Frequently Asked Questions
What is the difference between an investment committee memo and a deal summary?
A summary describes the asset and the numbers. A memo claims why the deal produces a return, names the assumptions that claim rests on, and states what would prove it wrong. The test is whether any sentence in the document could turn out to be false.
What should an investment memo template for real estate include?
Thesis, basis and pricing against verified comparables, the two or three load-bearing assumptions, the evidence behind each, a downside case tied to a specific miss, written kill criteria, and a recommendation with a conviction level. Sections that describe rather than claim belong in an appendix.
Who writes the kill criteria, and can they be changed?
The deal lead writes them before diligence begins and the committee approves them with the deal. They can be changed, but only in writing and with the committee's consent, because a threshold has value only if set before the answer was known.
Conclusion
The purpose of an investment committee memo is not to inform the committee. It is to expose a belief to attack while the capital is still uncommitted. A summary avoids that by design: it states unarguable things and lets a committee approve a deal without naming the numbers the outcome depends on. A falsifiable memo costs the author something, which is why it works. Name the thesis, rank the assumptions, quantify the miss, set the trigger, and sign it. The payoff comes twice: in a meeting that argues about the right variable, and five years later, when the firm opens the file and learns whether it was right for the reason it thought.