LIHTC pro forma: a $4.8 million gap with the whole fee deferred

An analyst at an institutional real estate investor asks Compound whether the capital stack closes on a new 100-unit affordable building before applying for tax credits. Compound picks a Qualified Census Tract in Miami-Dade County from HUD, Zillow and Census data and builds the 9% and 4% cases side by side.

Compound Research · Updated Nov 15, 2026 · 14 min read

Short answer

A LIHTC pro forma tests whether tax credit equity, a permanent loan and soft loans cover the cost of an affordable building. In Compound's model of a 100-unit building in a Miami-Dade Qualified Census Tract, the 9% case is $4.8 million short after deferring the whole developer fee, and the 4% case $5.3 million.

Try it now
On this page (12)
  1. 1. What goes into a LIHTC pro forma
  2. 2. Picking the site: the rent test and the Qualified Census Tract
  3. 3. LIHTC rent limits by AMI tier, net of a utility allowance
  4. 4. Development budget: hard cost from the state's limit, trended by the PPI
  5. 5. Eligible basis, the 30% basis boost and qualified basis
  6. 6. 9% vs 4% LIHTC: the annual credit and the tax credit equity
  7. 7. LIHTC sources and uses: the deferred developer fee as the balancing item
  8. 8. The 15-year pro forma: debt coverage and repaying the deferred fee
  9. 9. What closes the funding gap: boost, credit price, DSCR and soft sources
  10. 10. What public data cannot tell you about a tax credit deal
  11. 11. Our view: start a LIHTC model from the limit on the credit
  12. 12. FAQ

The capital stack on this building does not close under either tax credit. With the entire developer fee deferred, the 9% case is $4,807,375 short of a total development cost of $59,129,594, and the 4% case is $5,293,697 short of $59,999,594. Fifteen years of cash flow then repay only $2,065,480 of a $7,418,063 deferred fee. Count the fee still unpaid in Year 15 together with the remaining gap, and the 9% case is $10,159,958 short of paying every use and the 4% case is $10,766,280 short (our arithmetic).

You work in acquisitions or asset management at an institutional real estate investor, and a Low-Income Housing Tax Credit (LIHTC) development is on your desk. Before anyone writes an application, you need to know whether credit equity, a permanent loan and soft loans add up to the cost of the building, and how much developer fee has to stay in the deal. We asked Compound to build that test from public data for a site it picked itself. It ran on November 15, 2026 and returned one Excel workbook and a summary.

The site is real: a census tract on HUD’s 2026 list of Qualified Census Tracts in Miami-Dade County, Florida. The building is hypothetical. Its land price, soft costs, reserves, credit prices and soft loans are analyst assumptions, and the workbook marks each one. The gap is the output of those assumptions. It is neither a forecast for any property in the tract nor advice to buy, build or invest there.

What the workbook shows:

  • The state’s request limit sets the 9% credit. Qualified basis of $68,381,248 computes an annual credit of $6,154,312. The state’s request for applications limits a Miami-Dade request to $3,800,000, so $2,354,312 of computed credit a year cannot be used.
  • The 30% basis boost adds nothing to the 9% case. The gap before the deferred fee is $12,225,439 at every boost from 0% to 30%. In the 4% case the same boost is worth $5.4 million of equity (our arithmetic).
  • Restricted rents support a small loan. Year 1 net operating income of $773,216 at a 1.15x debt service coverage ratio and a 6.90% rate sizes a permanent loan of $8,867,499, which is 15% of cost (our arithmetic).
  • Cost is $591,296 per unit. The hard cost is the state agency’s published factor of $339,000 per unit, trended by a BLS construction price index to $358,474.
  • Three flags are on in both cases. The deferred fee is above 50% of the total fee, it is not repaid by Year 15, and a funding gap remains.
Prompt to try

“I am checking whether the capital stack closes on a new affordable apartment building before we apply for tax credits, and want a LIHTC pro forma anchored on public facts. Pick the site from the data: choose a county where the 60% of area median income rent in HUD's published LIHTC rent limits sits at least 15% below the HUD Fair Market Rent and the Zillow ZORI rent for the metro, and within it a census tract that HUD's published list names a Qualified Census Tract; show the tract's median household income and renter share from the Census ACS data. The subject is a 100-unit new-construction family building in that tract, with units at 30%, 50% and 60% of area median income. Set restricted rents from HUD's published rent limits net of a utility allowance. Set hard cost per unit from the state housing agency's published cost limits, brought to today with the BLS construction cost PPI, and set land, soft costs, reserves and the developer fee as analyst assumptions, with the fee held to the cap in the state's published Qualified Allocation Plan. Set the credit price per dollar of credit as an analyst assumption with a note. Price the permanent loan off today's 10-year Treasury yield plus a spread, sized on restricted-rent NOI at a 1.15x DSCR. Build: a switch between the 9% and the 4% credit; an eligible-basis worksheet that leaves out land, most reserves and permanent financing costs, applies the 30% basis boost for the tract, and multiplies by the applicable fraction to reach qualified basis; the annual credit and the ten-year credit total; a syndication panel that turns credits into equity at the credit price; sources and uses with each soft source on its own line and the deferred developer fee as the balancing item; and a 15-year operating pro forma that shows whether cash flow covers debt service and repays the deferred fee. Show the 9% and 4% cases side by side, and add sensitivity tables of the funding gap to the basis boost, credit price, DSCR floor and soft sources. Flag a deferred fee above half the total fee, a deferred fee not repaid by year 15, and a remaining gap under either credit. Use the stored public data wherever it has the figure, and web search only for what it lacks. Put every input on an Assumptions sheet with one column that says where it came from: a cited source, or "analyst assumption" for a figure I would set myself, with the reason for the value you chose. Every output must be a live formula that traces back to the Assumptions sheet. Add a Sources sheet listing each data source, what it gave and its date. Deliver one Excel workbook and a short summary of the result, the flags, and what I would need from the seller to firm it up. Cite every sourced figure.”

Try this prompt free →

That is the request as we sent it, without its closing line, which asks that no name or email be put in the workbook.

What goes into a LIHTC pro forma

A LIHTC pro forma answers one question: do the sources equal the uses. It has six linked parts: restricted rents, a development budget, an eligible-basis worksheet, the credit and its sale price, a loan sized on restricted income, and an operating forecast that repays the fee the developer deferred.

A market-rate development pro forma solves for a return on equity. A tax credit deal has almost no cash return to solve for. Rents are capped by law, so the income supports little debt, and most of the cost is paid by an investor who buys the tax credits.

Compound’s workbook follows that order on 11 sheets. The Assumptions sheet holds every input with a column for where it came from. Sourced inputs cite HUD, the Census Bureau, the Bureau of Labor Statistics, the Treasury, the Florida Housing Finance Corporation (FHFC) and Section 42 of the tax code. The rest say “Analyst assumption” and give a reason for the value. On this deal the sourced inputs are the rents, the hard cost, the credit rules and the base interest rate, and nearly everything else is yours to replace.

If you are buying an existing rent-restricted property, the affordable housing acquisition model is the closer fit. That page has no credits in it.

Picking the site: the rent test and the Qualified Census Tract

Miami-Dade County passes the request’s rent test by the widest margin of the five counties Compound compared: the two-bedroom rent at 60% of area median income (AMI) is 28.3% below HUD’s Fair Market Rent and 30.7% below Zillow’s metro rent. Inside the county, Compound chose Census Tract 30.04, which is on HUD’s 2026 Qualified Census Tract list and where 93.5% of occupied homes are rented.

The request asked for a county where the restricted rent sits at least 15% below both market measures.

Compound computed the 60% rent from HUD’s income limits, which it took from HUD’s website: the three-person limit of $73,560 in the Miami area, times 30%, divided by 12, gives $1,839 a month. HUD’s Fair Market Rent for a two-bedroom is $2,564, and the Zillow Observed Rent Index (ZORI) for the metro was $2,654 in September 2026. Both came from Compound’s stored public data.

Miami-Dade_LIHTC_Pro_Forma.xlsx
K6fx=IF(J6="Pass",MIN(G6,I6),-1)
BCDEFGHIJK
5CountyRent comparison metro (ZORI)60% AMI limit, 3-person60% AMI 2BR rentFMR 2BRRent below FMRZORI, metroRent below ZORIMeets testSmaller discount (passing only)
6Miami-Dade County, FLMiami-Fort Lauderdale-West Palm Beach metro$73,560$1,839$2,56428.3%$2,65430.7%Pass28.3%
7Los Angeles County, CALos Angeles-Long Beach-Anaheim metro$89,940$2,248$2,96424.2%$2,92923.2%Pass23.2%
8Bronx County, NYNew York-Newark-Jersey City metro$91,620$2,290$2,97122.9%$3,54635.4%Pass22.9%
9Broward County, FLMiami-Fort Lauderdale-West Palm Beach metro$68,400$1,710$2,32126.3%$2,65435.6%Pass26.3%
10Baltimore city, MDBaltimore-Columbia-Towson metro$72,360$1,809$2,0069.8%$1,9376.6%Failfail
12Selected county (largest discount among passing counties)Miami-Dade County, FL
13Required discount (Request)15%
15QUALIFIED CENSUS TRACT
16Census tract (GEOID)12086003004 (Census Tract 30.04, Miami-Dade County)
17On HUD 2026 Qualified Census Tract listYes
18Median household income (ACS 2020-2024)$38,929
19Occupied housing units2,614
20Renter-occupied units2,444
21Renter share93.5%
22Tract median income as share of HMFA 60% AMI limit, 3-person52.9%
SummaryAssumptionsSiteRentsBudgetBasisCreditsSources & UsesPro FormaSensitivitySources
The Site sheet, columns B to K, as Compound wrote them. We show the county test, the selected county and the tract block, and leave out the title rows and the note under the tract block. The formula bar shows the cell that keeps the smaller of a county's two discounts only when the county passes.

Four counties pass. Baltimore city fails: its 60% rent of $1,809 is 9.8% below the Fair Market Rent and 6.6% below ZORI. The selected county is a formula that takes the passing county with the largest value in the last column.

How far the 60% AMI rent sits below market, by county

The smaller of two discounts for a two-bedroom: against HUD's Fair Market Rent and against Zillow's metro rent index. The request required 15%

Miami-Dade County, FL28.3%Broward County, FL26.3%Los Angeles County, CA23.2%Bronx County, NY22.9%Baltimore city, MD6.6%Required by the request15.0%

Compound workbook, Site sheet, columns G, I and K. Income limits are HUD FY2026, Fair Market Rents are HUD FY2027, ZORI is September 2026.

The tract is typed onto the sheet with a note on how it was chosen: among Miami-Dade’s 164 tracts on the 2026 list, it has the most renter households. The Census ACS figures for 2020 to 2024 give it a median household income of $38,929, with 2,444 renter households out of 2,614 occupied homes. That income is 52.9% of the 60% limit for a three-person household.

LIHTC rent limits by AMI tier, net of a utility allowance

The 100 units average $1,557 a month in net rent, for $1,868,760 of annual rent at full occupancy. Each rent is 30% of the income limit for the unit’s imputed household size, divided by 12, less a utility allowance.

HUD publishes income limits, and the rent follows from them by the tax code’s formula. A unit is assumed to house 1.5 persons per bedroom, so a two-bedroom uses the three-person limit, and a one-bedroom or three-bedroom uses the average of two household sizes. Rent plus utilities cannot exceed 30% of that income, and the owner collects the rent net of the allowance.

UnitAMI tierUnitsIncome limit at imputed sizeMax gross rentUtility allowanceNet rentFair Market RentNet rent / FMR
1BR30%3$30,660$766$95$671$2,12132%
1BR50%5$51,100$1,277$95$1,182$2,12156%
1BR60%17$61,320$1,533$95$1,438$2,12168%
2BR30%5$36,780$919$125$794$2,56431%
2BR50%10$61,300$1,532$125$1,407$2,56455%
2BR60%35$73,560$1,839$125$1,714$2,56467%
3BR30%2$42,495$1,062$155$907$3,27728%
3BR50%5$70,825$1,770$155$1,615$3,27749%
3BR60%18$84,990$2,124$155$1,969$3,27760%
Total / average100$1,557

The unit mix is an analyst assumption: 25 one-bedrooms, 50 two-bedrooms and 25 three-bedrooms, with 10 units at 30% of AMI, 20 at 50% and 70 at 60% (the last two are our sums of the table). The 10 units at 30% meet the state’s minimum of 10.0% of units for extremely low-income households, and the Rents sheet checks the mix against it.

The utility allowances of $95, $125 and $155 are also assumptions. The source column says to replace them with the allowance approved by the county housing agency or by FHFC.

The 60% two-bedroom rents for $1,714, which is 67% of the Fair Market Rent, and the 30% units rent for about a third of it.

Development budget: hard cost from the state’s limit, trended by the PPI

Total development cost is $59,129,594 in the 9% case, or $591,296 per unit, and $59,999,594 in the 4% case. Hard cost is the one sourced line; land, soft costs and reserves are analyst assumptions, and the developer fee is 16.0% of a fee base, the cap in Florida’s Qualified Allocation Plan.

The request told Compound to set hard cost from the state housing agency’s published cost limits and bring it to today with a construction price index. Compound used FHFC’s hard cost factor for a mid-rise building in Broward, Miami-Dade and Palm Beach counties, $339,000 per unit as of April 2026. It multiplied that by the change in a BLS producer price index for construction inputs from April to October 2026, 162.244 to 171.564.

Miami-Dade_LIHTC_Pro_Forma.xlsx
C10fx=Assumptions!$C$79/Assumptions!$C$78
BCDE
49% case4% casePer unit, 9%
6ACQUISITION
7Land$4,500,000$4,500,000$45,000
8HARD COSTS
9FHFC hard cost factor per unit (Apr 2026)$339,000$339,000
10PPI trend factor, Oct 2026 / Apr 20261.0571.057
11Trended hard cost per unit$358,474$358,474
12Units100100
13Building and site hard cost$35,847,363$35,847,363$358,474
14Hard cost contingency$1,792,368$1,792,368$17,924
15Total hard costs ($)$37,639,731$37,639,731$376,397
26Total eligible soft costs ($)$7,543,165$7,543,165$75,432
33Total excluded soft costs ($)$900,000$1,650,000$9,000
35Operating deficit reserve$848,634$848,634$8,486
36Replacement reserve deposit$30,000$30,000$300
37Lease-up reserve$250,000$250,000$2,500
38Total reserves ($)$1,128,634$1,128,634$11,286
40Fee base: cost excluding land, operating deficit reserve and fee$46,362,896$47,112,896
41Fee rate (lower of requested and QAP cap)16.0%16.0%
42Developer fee ($)$7,418,063$7,538,063$74,181
43Total development cost ($)$59,129,594$59,999,594$591,296
SummaryAssumptionsSiteRentsBudgetBasisCreditsSources & UsesPro FormaSensitivitySources
The Budget sheet, columns B to E, as Compound wrote them. We show the land, hard cost, reserve and fee rows and the group totals, and leave out the individual soft cost lines and the 4% per-unit column. The formula bar shows the price index ratio that trends the state's April 2026 cost factor to October.

Compound’s answer is plain about that number. The FHFC factor is a ceiling, and the model uses it as the actual cost; real bids may come in lower. With a 5.0% contingency, hard cost is $37,639,731, which is 64% of the total (our arithmetic), so a contractor’s number is the input that would move this model most.

The budget is grouped by how the tax code treats each line. Basis-eligible soft costs total $7,543,165 and include construction interest and architecture. Costs excluded from basis are $900,000 in the 9% case and $1,650,000 in the 4% case, which adds $750,000 of bond issuance cost. Reserves are $1,128,634, and the operating deficit reserve inside that figure is a formula on six months of operating expenses and debt service.

Eligible basis, the 30% basis boost and qualified basis

Eligible basis before the boost is $52,600,960 in the 9% case. The tract’s 30% boost raises it to $68,381,248, and an applicable fraction of 100.0% makes that the qualified basis.

Eligible basis is the depreciable cost of the residential building. The worksheet starts from total development cost and takes out three groups: land of $4,500,000, the permanent financing, syndication, agency and marketing costs, and all reserves. A check row rebuilds the same figure as hard costs plus eligible soft costs plus developer fee and shows a difference of 0.

Miami-Dade_LIHTC_Pro_Forma.xlsx
C15fx=C9*(1+C14)
BCD
49% case4% case
5Total development cost$59,129,594$59,999,594
6Less: land($4,500,000)($4,500,000)
7Less: permanent financing, syndication, agency and marketing costs($900,000)($1,650,000)
8Less: reserves($1,128,634)($1,128,634)
9Eligible basis before boost ($)$52,600,960$52,720,960
10Check: hard costs + eligible soft costs + developer fee$52,600,960$52,720,960
11Difference$0$0
12QCT BASIS BOOST
13Tract on HUD 2026 QCT list (1 = yes)11
14Basis boost applied30%30%
15Eligible basis after boost ($)$68,381,248$68,537,248
16APPLICABLE FRACTION
17Low-income units100100
18Total units100100
19Unit fraction100.0%100.0%
20Applicable fraction (unit fraction; floor-space fraction taken as equal)100.0%100.0%
21Qualified basis ($)$68,381,248$68,537,248
SummaryAssumptionsSiteRentsBudgetBasisCreditsSources & UsesPro FormaSensitivitySources
The Basis sheet, columns B to D, rows 4 to 21, as Compound wrote them. The formula bar shows the cell that applies the basis boost, which is 30% only while the tract flag on the Assumptions sheet is 1.

A building in a Qualified Census Tract may multiply its eligible basis by 1.30x. The workbook reads the multiplier from the Assumptions sheet, where it cites Section 42(d)(5)(B), and applies it only if the tract flag is 1.

The applicable fraction is the low-income share of the building. Every unit here is restricted, so the fraction is 100.0%. The row label states the one shortcut: the floor-space fraction is taken as equal to the unit fraction.

The developer fee is inside basis, so a larger fee raises the computed credit. That is why the state caps the fee.

9% vs 4% LIHTC: the annual credit and the tax credit equity

The 9% case raises $33,436,656 of equity and the 4% case raises $23,300,334. The 9% credit is set by the state’s $3,800,000 request limit, and the 4% credit is set by qualified basis.

Qualified basis times the applicable percentage is the annual credit. The investor receives it for 10 years and pays a price per dollar of credit. The two cases differ at each step.

Miami-Dade_LIHTC_Pro_Forma.xlsx
C9fx=MIN(C7,C8)
BCDE
49% case4% caseActive case
5Qualified basis$68,381,248$68,537,248$68,381,248
6Applicable percentage9.0%4.0%9.0%
7Annual credit, computed$6,154,312$2,741,490$6,154,312
8Annual credit cap (9%: RFA maximum request; 4%: none)$3,800,000$2,741,490$3,800,000
9Annual credit ($)$3,800,000$2,741,490$3,800,000
10Credit period (years)101010
11Ten-year credit total ($)$38,000,000$27,414,899$38,000,000
12Investor share of credits99.99%99.99%99.99%
13Credits purchased by investor$37,996,200$27,412,158$37,996,200
14SYNDICATION
15Credit price per $1 of credit$0.88$0.85$0.88
16LIHTC equity ($)$33,436,656$23,300,334$33,436,656
17Equity per $1 of annual credit$8.80$8.50$8.80
18Computed credit above the RFA cap (not usable)$2,354,312$0$2,354,312
204% BOND TEST
21Tax-exempt bonds$20,000,000
22Aggregate basis (total cost less reserves)$58,870,960
23Bonds as share of aggregate basis34.0%
24Required minimum25.0%
25Test resultMet
SummaryAssumptionsSiteRentsBudgetBasisCreditsSources & UsesPro FormaSensitivitySources
The Credits sheet, columns B to E, as Compound wrote them, without the title rows and the last row. The Active case column follows the 9% or 4% switch on the Assumptions sheet. The formula bar shows the cell that takes the smaller of the computed credit and the cap.

In the 9% case the computed credit is $6,154,312 a year. FHFC’s request for applications, which Compound found by web search, limits a Miami-Dade request to $3,800,000. The model takes the smaller figure and reports the difference, $2,354,312, on a row labeled “not usable”. At $0.88 per dollar and a 99.99% investor share, that unused credit would have bought about $20.7 million of equity over the ten years (our arithmetic).

Annual credit: what basis computes and what the deal can use

Dollars per year for ten years. The 9% credit is competitive and limited by the state's request for applications. The 4% credit comes with tax-exempt bonds

9% case: credit the qualified basis computes$6.15M9% case: maximum request in the state's RFA$3.80M9% case: computed credit above the cap$2.35M4% case: credit the qualified basis computes$2.74M

Compound workbook, Credits sheet, rows 7, 8 and 18. The cap is from FHFC RFA 2026-203.

The 4% credit comes with tax-exempt bond financing instead of the state’s competitive allocation, and it has a different test: bonds must finance at least 25.0% of the building’s aggregate basis. The model assumes $20,000,000 of bonds against an aggregate basis of $58,870,960, a 34.0% share, and reports the test as met.

Both credit prices are analyst assumptions: $0.88 for the 9% credit and $0.85 for the 4%. The note beside them says to confirm with letters of intent from two syndicators. Each $0.02 of price is worth $759,924 of 9% equity (our arithmetic from the model’s price table), so the price would have to reach about $1.01 to close the 9% gap without any other change (our arithmetic). That is well outside the range the workbook’s note describes.

LIHTC sources and uses: the deferred developer fee as the balancing item

The permanent loan is $8,867,499 in both cases, and soft loans are $4,600,000 in the 9% case and $15,000,000 in the 4% case. That leaves a gap of $12,225,439 and $12,831,761 before any fee is deferred, and each gap is larger than the whole developer fee.

Year 1 net operating income of $773,216 divided by 1.15x allows $672,362 of annual debt service. The rate is the 10-year Treasury yield of 5.15% at the November 13, 2026 close plus an assumed spread of 1.75%, or 6.90%, on a 35-year amortization. The mortgage constant is 0.0758, and debt service divided by the constant is the loan.

Miami-Dade_LIHTC_Pro_Forma.xlsx
C39fx=MIN(C33,MAX(0,C38))
BCDE
49% case4% caseActive case
5PERMANENT LOAN SIZING
6Year 1 net operating income (restricted rents)$773,216$773,216$773,216
7Minimum DSCR1.15x1.15x1.15x
8Maximum annual debt service$672,362$672,362$672,362
910-year Treasury yield5.15%5.15%5.15%
10Spread1.75%1.75%1.75%
11Interest rate6.90%6.90%6.90%
12Amortization (years)353535
13Annual mortgage constant0.07580.07580.0758
14Permanent loan amount ($)$8,867,499$8,867,499$8,867,499
15Annual debt service$672,362$672,362$672,362
16SOURCES
17Permanent first mortgage$8,867,499$8,867,499$8,867,499
18LIHTC equity$33,436,656$23,300,334$33,436,656
19Miami-Dade County Surtax loan$3,000,000$5,000,000$3,000,000
20FHFC ELI gap loan$600,000$1,000,000$600,000
21FHFC SAIL loan$0$8,000,000$0
22Miami-Dade County HOME loan$1,000,000$1,000,000$1,000,000
23Deferred developer fee (balancing item)$7,418,063$7,538,063$7,418,063
24Remaining funding gap (unfunded)$4,807,375$5,293,697$4,807,375
25Less: excess sources (credits or soft funds to be cut)$0$0$0
26Total sources ($)$59,129,594$59,999,594$59,129,594
34Total uses ($)$59,129,594$59,999,594$59,129,594
36GAP AND DEFERRED FEE
37Total soft sources$4,600,000$15,000,000$4,600,000
38Funding gap before deferred fee ($)$12,225,439$12,831,761$12,225,439
39Deferred developer fee$7,418,063$7,538,063$7,418,063
40Remaining funding gap after full fee deferral ($)$4,807,375$5,293,697$4,807,375
41Excess sources$0$0$0
42Deferred fee as share of total developer fee100%100%100%
43Check: total sources less total uses$0$0$0
SummaryAssumptionsSiteRentsBudgetBasisCreditsSources & UsesPro FormaSensitivitySources
The Sources & Uses sheet, columns B to E, as Compound wrote them. We show the loan sizing, every source, the total uses and the gap block, and leave out the six use lines above the total. The formula bar shows the deferred fee: the gap, limited to the fee itself and never below zero.

Each soft source sits on its own line, as the request asked: a county surtax loan, two FHFC loans and a county HOME loan. All four amounts are analyst assumptions with a reason in the source column, and none is a commitment. The 4% case assumes $10.4 million more of them than the 9% case (our arithmetic), which nearly offsets its $10.1 million of lower equity (our arithmetic).

The deferred developer fee is the balancing item. One formula defers as much fee as the gap needs, up to the entire fee, and a second reports anything left over as a remaining funding gap. In both cases the deferral reaches 100% of the fee and the leftover is positive. Total sources equal total uses because the sheet prints the unfunded gap as a source line with that label.

Where $59.1 million of cost is paid from in the 9% case

Dollars. The deferred fee is paid to the developer later from cash flow. The remaining gap has no source in the model

LIHTC equity$33.4MPermanent first mortgage$8.87MSoft sources (three loans)$4.60MDeferred developer fee$7.42MRemaining funding gap (unfunded)$4.81M

Compound workbook, Sources & Uses sheet, rows 17 to 24, 9% case column.

Credit equity pays for 57% of the 9% building and 39% of the 4% building (our arithmetic). In a real estate pro forma for a market-rate asset, debt is usually the largest source. Here the loan is the third-largest line in the 4% case, behind the equity and the soft loans.

The 15-year pro forma: debt coverage and repaying the deferred fee

Net operating income covers debt service in every year, with coverage rising from 1.15x to 1.25x. Cash flow after debt service totals $2,065,480 over 15 years, which is 28% of the deferred fee in the 9% case (our arithmetic).

The forecast grows rents at 2.00% a year and expenses at 3.00%, both analyst assumptions. Vacancy is 5.0%. Operating costs are entered per unit: $4,800 for operations, $2,400 for insurance, $1,800 for real estate taxes and $350 for replacement reserves, plus a management fee of 5.0% of income. Year 1 expenses are 57% of effective gross income (our arithmetic).

Miami-Dade_LIHTC_Pro_Forma.xlsx
C24fx=-MIN(C22,C23)
BCDGLQ
4YearYear 1Year 2Year 5Year 10Year 15
5OPERATIONS
6Gross potential rent$1,868,760$1,906,135$2,022,806$2,233,341$2,465,789
7Other income$24,000$24,480$25,978$28,682$31,667
8Vacancy and collection loss($94,638)($96,531)($102,439)($113,101)($124,873)
9Effective gross income ($)$1,798,122$1,834,084$1,946,345$2,148,922$2,372,584
15Total operating expenses ($)($1,024,906)($1,054,754)($1,149,668)($1,327,409)($1,532,901)
16Net operating income ($)$773,216$779,330$796,677$821,513$839,683
17Permanent loan debt service($672,362)($672,362)($672,362)($672,362)($672,362)
18Cash flow after debt service ($)$100,854$106,969$124,315$149,152$167,322
19DSCR1.15x1.16x1.18x1.22x1.25x
20NOI covers debt serviceYesYesYesYesYes
21DEFERRED DEVELOPER FEE: 9% CASE
22Beginning balance$7,418,063$7,317,209$6,978,605$6,305,071$5,519,905
23Cash flow available for fee$100,854$106,969$124,315$149,152$167,322
24Repayment($100,854)($106,969)($124,315)($149,152)($167,322)
25Ending balance$7,317,209$7,210,241$6,854,290$6,155,920$5,352,583
26DEFERRED DEVELOPER FEE: 4% CASE
30Ending balance$7,437,209$7,330,241$6,974,290$6,275,920$5,472,583
SummaryAssumptionsSiteRentsBudgetBasisCreditsSources & UsesPro FormaSensitivitySources
The Pro Forma sheet as Compound wrote it, showing Years 1, 2, 5, 10 and 15 of the 15 yearly columns. We leave out the five expense lines above their total and most of the 4% fee block, whose ending balance is the last row. The formula bar shows the fee repayment: the smaller of the balance owed and the year's available cash.

Expenses grow faster than rents, so net operating income rises slowly, from $773,216 in Year 1 to $839,683 in Year 15. Debt service is level at $672,362. Cash flow after debt service is $100,854 in Year 1 and $167,322 in Year 15, and the model sends 100% of it to the deferred fee.

The 9% balance falls from $7,418,063 to $5,352,583 at the end of Year 15, and the 4% balance from $7,538,063 to $5,472,583. The Summary sheet carries six flags as formulas:

Flag9% case4% case
Deferred fee above 50% of total feeFLAGFLAG
Deferred fee not repaid by Year 15FLAGFLAG
Remaining funding gapFLAGFLAG
Cash flow fails to cover debt service in any yearOKOK
4% bond test (25% of aggregate basis)n/aOK
ELI set-aside at or above FHFC minimumOKOK

The three flags the request named are on in both cases. The other three are checks Compound added, and they pass.

What closes the funding gap: boost, credit price, DSCR and soft sources

Only one cell in the four sensitivity tables shows no remaining gap: the 4% case with soft sources at 150% of the assumed amounts, or $22,500,000. Even there $5,331,761 of the developer fee stays deferred, which is 71% of the fee (our arithmetic) and above the 50% flag. In the 9% case no single change in the tables closes the gap.

These rows come from the four tables on the Sensitivity sheet:

Change from the base caseRemaining gap, 9%Remaining gap, 4%
Base case$4,807,375$5,293,697
No basis boost (0%)$4,807,375$10,670,697
Basis boost of 20%$4,807,375$7,086,031
Credit price two steps lower ($0.84 and $0.81)$6,327,223$6,390,184
Credit price two steps higher ($0.92 and $0.89)$3,287,527$4,197,211
DSCR floor of 1.05x$3,994,869$4,481,191
DSCR floor of 1.25x$5,489,881$5,976,203
Soft sources at 50% of the assumed amounts$7,107,375$12,793,697
Soft sources at 125% of the assumed amounts$3,657,375$1,543,697
Soft sources at 150% of the assumed amounts$2,507,375$0

The boost rows show the request limit at work. Without any boost, 9% of eligible basis is still $4,734,086 (our arithmetic), above the $3,800,000 limit, so the 9% equity is the same with or without the tract. The tract is worth money only in the 4% case, where the remaining gap is $10,670,697 without it.

Remaining funding gap after deferring the whole developer fee

Dollars. Each bar changes one input from the base case: credit price two steps higher, DSCR floor at 1.05x, or soft sources at 150%

9% case, base$4.81M9%, credit price $0.92$3.29M9%, DSCR floor 1.05x$3.99M9%, soft sources at 150%$2.51M4% case, base$5.29M4%, credit price $0.89$4.20M4%, DSCR floor 1.05x$4.48M4%, soft sources at 150%$0.00M

Compound workbook, Sensitivity sheet, tables 2, 3 and 4, remaining gap columns.

A lower coverage floor helps less than you might expect. At 1.05x the loan grows to $9,712,023, but the operating deficit reserve grows with debt service, and the 9% gap falls only to $3,994,869.

What public data cannot tell you about a tax credit deal

Public data gives you the rent limits, the tract’s status, the state’s cost ceiling and the rules. A land price, a construction bid, a credit price and a funding commitment come from other people, and those four decide whether this stack closes.

Compound’s answer ends with what you would need from the seller of the site to firm up the model:

  • The land contract price, survey, title, Phase I and appraisal. Land is $4,500,000 in the model, about $45,000 per unit, with a note to replace it with the contract price. Land is outside basis, so it earns no credit.
  • Zoning and entitlements, allowed density and parking. The model assumes 100 units fit.
  • Site plan, building type and contractor bids. A real bid replaces the largest line in the budget.
  • Impact fee and utility connection fee estimates. Permits and impact fees are $15,000 per unit as an assumption.
  • Flood zone and insurance quotes. Insurance is $2,400 per unit a year, an assumption a broker’s quote should replace.
  • Property tax history and exemption status. Taxes are $1,800 per unit.
  • Existing city or county funding commitments or land-use restrictions on the parcel.
  • Confirmation that every parcel is inside tract 12086003004. The boost depends on it.

Three inputs come from other parties. The syndicator sets the credit price. The county housing agency or FHFC sets the utility allowance. The soft lenders set their amounts and terms. The answer also notes that the 4% case applies the 9% fee cap and that Florida’s separate rule for bond deals should be checked.

Our view: start a LIHTC model from the limit on the credit

On a 9% deal in a state with a request limit, the credit is a fixed number of dollars, and every cost above what it supports falls on the fee and the soft loans. Read that limit before you build the basis worksheet.

Most LIHTC models are built bottom-up: budget, basis, boost, credit. This workbook was built that way too, as the request asked, and its most useful output is the row that shows the bottom-up credit running $2,354,312 a year above the request limit. Once the limit binds, the boost stops mattering, and the deal is decided by cost per unit against a fixed equity amount of $33,436,656.

That makes the 9% case a cost problem. A remaining gap of $4,807,375 is 8.1% of total cost, or $48,074 per unit (both our arithmetic). FHFC’s requests for applications also limit total development cost per unit by building type, so check $591,296 per unit against the limit in the current request before you apply. A contractor’s bid below the agency’s ceiling, cheaper land or a larger county loan could each cover part of it, and none of those is in public data.

The 4% case is a soft-money problem. Its credit has no request limit but is smaller, and it closes only with $22,500,000 of soft loans, $7.5 million above the base assumption (our arithmetic). At that level $5,331,761 of the fee is still deferred, and 15 years of cash flow repay $2,065,480 of it. We would not present either case as fundable from this model. We would use it to decide which calls to make first: a contractor, a syndicator and the county.

LIHTC pro forma, 9% and 4% cases (Excel)Compound’s workbook from this run: a Summary with both cases and six flags; the Assumptions sheet with a source or a reason for every input; the Site rent test for five counties; restricted Rents by unit type and income tier; the development Budget; the eligible Basis worksheet; Credits and syndication; Sources & Uses; the 15-year Pro Forma; four Sensitivity tables; and the Sources list.Download · Free

FAQ

What is a LIHTC pro forma?

It is a financial model for a Low-Income Housing Tax Credit development. It sets rents from HUD income limits, builds a development budget, computes eligible and qualified basis, turns the annual credit into investor equity, sizes a permanent loan on restricted income, and lays out sources and uses. A 15-year operating forecast then tests debt coverage and repayment of any deferred developer fee.

What is the difference between 4% and 9% LIHTC?

The 9% credit is awarded by competition under a state’s allocation plan. The 4% credit comes with tax-exempt bond financing. In this model the 9% credit is limited to $3,800,000 a year and raises $33,436,656 of equity, and the 4% credit of $2,741,490 a year raises $23,300,334.

How do you calculate eligible basis and qualified basis?

Start with total development cost and remove land, reserves, and permanent financing, syndication and marketing costs. That is eligible basis: $52,600,960 here. Multiply by 1.30x if the building is in a Qualified Census Tract, then by the applicable fraction, the low-income share of the building. Here the fraction is 100.0%, so qualified basis is $68,381,248.

How is tax credit equity calculated?

Multiply the annual credit by the 10-year credit period, by the investor’s share of the partnership, and by the price the investor pays per dollar of credit. In the 9% case that is $3,800,000 times 10, times 99.99%, times $0.88, which gives $33,436,656. The credit price is an analyst assumption in this model.

What is a deferred developer fee in a LIHTC deal?

It is the part of the developer fee that is left in the deal as a source and paid later from cash flow. Models treat it as the balancing item between sources and uses. Here the whole fee of $7,418,063 is deferred in the 9% case, and cash flow repays it only down to $5,352,583 by Year 15, so the workbook flags it twice.

CoatueCentri ConsultingAvra783 Capital PartnersOctahedron CapitalArix ResearchCoatueCentri ConsultingAvra783 Capital PartnersOctahedron CapitalArix ResearchCoatueCentri ConsultingAvra783 Capital PartnersOctahedron CapitalArix Research

See this workflow on your own deal

Try it now