How Housing Lotteries Actually Work

A housing lottery is a random drawing used to decide the order in which applicants are considered for a limited housing resource: a unit, a voucher, or a position on a waiting list.

That is the whole concept. What makes lotteries confusing is not the idea but the variation: agencies run them under different programs, with different rules, at different scales, and the word "lottery" gets applied to processes that work quite differently from each other.

This post walks through how they actually work, in plain terms. It is written for people who administer them, and for the boards, applicants, and partners who need to understand what happened.

When agencies use a lottery

A lottery becomes the right tool at a specific moment: when demand exceeds supply and there is no defensible basis for preferring one applicant over another.

Four situations produce that condition regularly.

Reopening a closed waiting list. This is the largest and most visible case. A public housing authority closes its waiting list because it is years long, then reopens it. Applications arrive by the thousands in a short window. Every applicant met the same criteria and applied inside the same period, so first-come-first-served would reward nothing except internet speed and time zone. A drawing establishes the order instead.

Filling units in a new affordable development. A building comes online with a fixed number of income-restricted units and far more qualified applicants than units. LIHTC properties and inclusionary developments run these constantly.

Municipal inclusionary programs. A city ordinance requires a developer to set aside affordable units. Local rules typically require a public, random selection process, often with preferences for current residents or municipal employees.

Special-purpose programs. Voucher set-asides, VASH allocations, and similar targeted programs where the eligible pool exceeds the available slots.

What unites these is that the agency is not choosing who deserves housing. Everyone in the pool already qualifies. The lottery decides only the order in which qualified applicants are reached.

The basic mechanics

At its simplest, a housing lottery takes a list and returns the same list in a random order.

INPUT                          OUTPUT
Application 4471               1.  Application 8823
Application 8823               2.  Application 2109
Application 2109        →      3.  Application 6640
Application 6640               4.  Application 4471
Application 5512               5.  Application 5512

Everything else (weighting, preference tiers, unit matching) is a modification to that core operation.

Two properties matter more than anything else about how the reordering is done.

It has to be genuinely unpredictable. A drawing that can be predicted in advance can be gamed, and one that can be reproduced by someone who knows the starting conditions is not random in the sense that matters. This is why serious lottery processing uses cryptographically secure random number generation (CSPRNG) rather than the ordinary random functions built into spreadsheets and programming languages, which are designed to be fast and repeatable rather than unpredictable.

It has to be unbiased. Every entry needs an equal chance at every position, with no subtle drift toward the beginning or end of the input list. Naive shuffling methods introduce exactly that kind of drift, and it is invisible unless you go looking for it.

Weighting: giving an applicant more chances

Many programs give certain applicants better odds without guaranteeing them a place. This is called weighting, and it is the most commonly misunderstood mechanism in housing lotteries.

Weighting works by giving an applicant multiple entries in the same drawing. An applicant with a weight of 3 has three chances instead of one. They are three times as likely as a weight-1 applicant to be drawn early, and they are not guaranteed to be.

A worked example with three applicants:

Six entries go into the drawing. Casey holds three of them and is three times as likely as Avery to come out first. But Avery can still place ahead of everyone. A weight of 3 is not a promise, it is a probability.

This distinction causes real friction when it is not explained clearly up front. Applicants who understand they received "extra weight" often hear it as "priority," and are genuinely surprised to place behind someone with fewer chances. Saying plainly that weighting improves odds rather than guaranteeing position prevents a category of complaint that is otherwise almost guaranteed.

Keep weight ranges small. Most agencies use 1–3 or 1–5. Very large weights rarely reflect an actual policy and become difficult to defend, because at a weight of 50 the applicant is not really in a lottery anymore.

Weighting versus preference tiers

These two mechanisms get conflated constantly, and they behave very differently.

A preference tier is not a weight. In a tiered drawing, every applicant in the higher tier is placed ahead of every applicant in the lower tier, with randomization happening inside each tier. If your policy says veterans are served before non-veterans, that is a tier. If it says veterans get an additional chance, that is a weight. The two produce very different outcomes, and using the wrong word in a public notice creates a real problem when the results come out.

What actually happens, step by step

Whether run internally or by an outside processor, a well-run drawing follows the same sequence.

1. Prepare the applicant data. One row per entry, with a consistent identifier. Most agencies use an application or ticket number rather than a name, so results can be published without exposing personal information.

2. Validate before running anything. Check for duplicate identifiers, missing weight values, malformed rows, and structural problems. Errors caught here are inconvenient; errors caught after the drawing are a re-run at best.

3. Lock the inputs. Once the drawing is scheduled, the applicant data and settings should be frozen. A drawing whose inputs could still change is not a drawing anyone can defend.

4. Run the drawing. Entries are drawn one at a time in a controlled environment, each receiving its own timestamp.

5. Generate the record. Full results, a public-safe version, and a processing log, created at the moment of completion rather than assembled afterward.

6. Handle the data responsibly. Applicant information should exist for as short a time as the process allows. Anything retained is something that can be exposed later.

7. Distribute. Results to staff, a censored version for public and board use, and documentation filed for the retention period your program requires.

Common questions

How many applicants can a housing lottery include? There is no inherent limit. Unit and property lotteries commonly run from a few dozen to a few hundred entries. Waiting list openings at large agencies can produce tens of thousands of applications at once. The mechanics are identical; only the scale changes.

Is a lottery required by HUD? No. Federal regulations for the Housing Choice Voucher program do not prescribe a lottery. 24 CFR § 982.202(d) requires that a PHA's admission policy state the system of preferences it uses. A lottery is a permitted local method. The requirement is that your written policy describes what you do and that you do what it describes.

Can a lottery consider income, family size, or household composition? Selection order cannot be based on protected characteristics. 24 CFR § 982.202(b) prohibits basing admission on age, race, color, religion, sex, national origin, disability, or familial status. Eligibility screening and unit matching are separate steps from the random ordering itself, and it is worth keeping them clearly separate in both your process and your documentation.

What is the difference between a lottery and a waiting list? A lottery is an event that produces an order. A waiting list is the ongoing record that order feeds into. Agencies often run a lottery to establish the initial sequence of a newly opened list, then work down that list over months or years.

Can results be published? Yes, with personally identifiable information removed. Publishing ticket or application numbers alongside draw positions lets every applicant find their own result without exposing anyone's identity, which is one of several reasons to draw by application number rather than by name.

What if two applicants have the same name or identifier? Draw by a unique identifier, not by name. Names are not unique and never have been. Application or ticket numbers are, and they double as the safe public reference.

Why the process gets scrutinized

Housing lotteries attract more attention than most administrative procedures, for a straightforward reason: the stakes are unusually high and the outcome is unusually visible.

An applicant who places at position 8 gets housed. An applicant at position 400 may wait years. Same qualifications, same application window, different outcome, determined by a process most applicants never see.

That combination invites questions, and the questions are legitimate. An agency that can explain its method clearly, show that the drawing was genuinely random, and produce a record demonstrating the results have not been altered answers them quickly. An agency that cannot ends up defending its integrity instead of its process, which is a much harder conversation.

The mechanics of a housing lottery are simple. The documentation around it is what determines whether the result holds up, and that is best settled before the first drawing rather than after the first challenge.

Lenora Castrellon is the founder of AuditDraw, which processes housing lotteries as a neutral third party for housing authorities, municipal housing departments, and affordable-housing organizations, and has been running real housing lotteries since 2019.

Nothing here is legal advice. Program rules vary by jurisdiction and funding source; your counsel and your funder's requirements govern.