Underwriting a deal end-to-end

Underwriting a deal means building the complete financial model that says: at this price, with this financing, this property produces this IRR for this investor under these assumptions, and these are the assumptions you should worry about. Modules 2-5 supplied the components. This module assembles them into a single coherent model on a worked example.

The six layers of an underwriting model

• Purchase assumptions — price, transaction costs (legal, stamp duty, broker), reserves at close, the all-in 'basis' the equity needs to fund.
• Financing — loan amount, interest rate, amortisation, fees, refi or sale at exit, debt service schedule.
• Operating pro forma — year-by-year NOI bridge from GPI down to NOI, capex schedule, leasing-cost schedule, unlevered cash flow.
• Exit assumptions — exit cap rate, year-of-sale NOI used for direct cap, transaction costs at sale, net proceeds.
• Returns layer — cash-on-cash by year, DSCR by year, unlevered IRR, levered IRR, equity multiple, peak equity exposure.
• Sensitivity and stress — sensitivity tables on the variables that move the answer, stress tests for the bad scenarios.

Worked underwriting: a Kileleshwa apartment block

PROPERTY:    24-unit residential, Kileleshwa
HOLD:        5 years, planned exit by sale
DATE:        Year 0 = today

─── PURCHASE ─────────────────────────────────────────
Purchase price                           KES   240,000,000
Transaction costs (legal, stamp duty 4%)       9,600,000
Initial reserves (3 months opex)               1,800,000
──────────────────────────────────────────────────────
Total basis                              KES   251,400,000

─── FINANCING ────────────────────────────────────────
Senior debt at 65% LTV  (on price)             156,000,000
Mortgage rate                                       13.0%
Amortisation                            20 years monthly
Monthly payment ~                              1,828,000
Annual debt service ~                         21,936,000

Equity check at close                         95,400,000

─── OPERATING PRO FORMA  (KES m)  ────────────────────
                          Yr 1   Yr 2   Yr 3   Yr 4   Yr 5
Gross potential income    33.0   34.8   36.7   38.7   40.8
Vacancy & credit loss     (3.3)  (2.8)  (2.6)  (2.3)  (2.5)
Other income               1.6    1.7    1.8    1.9    2.0
Effective gross income    31.3   33.7   35.9   38.3   40.3
Operating expenses       (11.0) (11.5) (12.1) (12.7) (13.3)
─────────────────────────────────────────────────────
Net operating income      20.3   22.2   23.8   25.6   27.0
Capex (reserve)           (2.4)  (2.4)  (2.4)  (2.4)  (2.4)
Unlevered cash flow       17.9   19.8   21.4   23.2   24.6
Debt service             (21.9) (21.9) (21.9) (21.9) (21.9)
Levered cash flow         (4.0)  (2.1)  (0.5)   1.3    2.7

DSCR                       0.93   1.01   1.09   1.17   1.23

─── EXIT (Year 5) ────────────────────────────────────
Forward NOI year 6                              28.4
Exit cap rate                                   8.5%
Gross sale price                              334.1
Sale costs (broker 3%, legal 1%)             (13.4)
Net sale proceeds                             320.7
Loan balance at exit (after 5 yrs amort)    (128.0)
Net equity proceeds at sale                   192.7

─── RETURNS ──────────────────────────────────────────
Cash-on-cash, stabilised (Yr 5)             2.8%
Unlevered IRR                              ~11.3%
Levered IRR                                ~13.9%
Equity multiple                              2.0×

Reading this deal

Three things jump out. First, the building is cash-flow negative to equity for the first three years — vacancy is high during lease-up and the debt service swallows everything else. The sponsor needs working-capital reserves to bridge this, and the LP needs to be comfortable that they're not getting paid out of operations for a while. Second, DSCR is below the lender's 1.20× covenant in years 1-2; either the lender accepts a covenant breach (with a cure mechanism), or the deal needs an interest-only initial period, or it doesn't get financed at this leverage. Third, the levered IRR is 13.9% — about 250bp above the unlevered. That positive spread says financing at 13% is roughly economic against a property earning ~11% unlevered, mostly because of the leverage applied to the appreciation between the going-in and going-out cap rate.

The two sensitivities that matter

Build a 2-D sensitivity table with exit cap rate on one axis and rent growth on the other. Almost every deal's IRR is mostly determined by these two variables; everything else is rounding.

LEVERED IRR SENSITIVITY  —  Kileleshwa example

              Exit cap rate
              7.5%    8.0%    8.5%    9.0%    9.5%
            ───────────────────────────────────────
Rent  3%      14.9%   13.0%   11.3%    9.7%    8.2%
growth 5%     17.4%   15.4%   13.9%   12.0%   10.6%
      7%      20.0%   17.9%   16.0%   14.4%   12.9%

Two things to read from this table. First, the spread from the worst cell (8.2%) to the best (20.0%) is 1,180 basis points. The deal isn't really 'a 14% IRR deal' — it's a deal whose IRR depends almost entirely on what happens with cap rates and rent growth over the next five years. Second, exit cap rate compression by 100bp (8.5 → 7.5) is worth approximately the same as 200bp of additional annual rent growth. That tells you where to focus your market diligence: the cap rate at exit matters as much as everything else combined.

Stress tests

Sensitivity tables explore variations of the base case. Stress tests ask what breaks the deal. The standard real-estate stresses are:

• Vacancy shock. Hold rent flat, push vacancy to 25% in years 1-2, recover to 10% by year 5. Does DSCR ever cure? Is the equity wiped out before stabilisation?
• Interest-rate shock. Push the mortgage rate up 300bp at refi (year 5). Does the sponsor's exit assumption survive a higher-rate world?
• Cap-rate decompression. Exit at the going-in cap rate plus 200bp (10.5% instead of 8.5%). What's the IRR? Is it still positive?
• Concentration loss. Lose the largest tenant in year 2. How long does re-leasing take, and what does the lost income do to the multi-year IRR?
• Construction / capex overrun. The KES 2.4m/year reserve assumption proves light; actual annual capex runs 5m. Does the deal still clear cost?

A deal that survives the base case but breaks under any of these stresses is a fragile deal. A deal that survives all five is a robust deal. Most institutional real-estate investors will reject any deal whose levered IRR turns negative under the 'cap-rate decompression + flat rents' stress — that's a deal whose return is entirely a directional bet on rates and cap rates, not on operating execution.

Going-in vs going-out cap rate — where alpha lives

The going-in cap rate is the cap rate you bought at: NOI year 1 divided by purchase price. The going-out cap rate is the cap rate the market will use to value the building at exit. The spread between them determines whether you are riding cap-rate compression (going-in > going-out, the market is more aggressive at your exit) or cap-rate decompression (going-in < going-out, the market is more cautious at your exit).

Most real-estate alpha — the excess return above what NOI growth alone would produce — comes from cap-rate compression. You buy a 9% cap rate today, hold for 5 years while rents grow 5% per year, and exit at a 7.5% cap rate because the market has decided this property type or location is more desirable than it was. The combination of NOI growth and cap-rate compression produces equity returns of 20%+ — the famous 'real estate makes you rich slowly and then suddenly'. The danger is the symmetric scenario: you buy at 7.5% and exit at 9% because the market has cooled. Same NOI growth, completely different outcome. Honest underwriting always shows the IRR at the same going-out cap rate as going-in (the 'no cap-rate compression' case) so the investor can see how much of the projected return depends on a favourable cap-rate move.