Identifying bottlenecks in the housing production chain
8 min read
Lesson objectives
- Distinguish observed delay from its cause and rank supply constraints by time, impact and tractability.
Please sign in to use all the tools
Once you sign in, the reading tools are yours: larger or smaller text, typeface, line spacing, focus mode, highlighting and bold, and notes saved to your account and gathered in My study.
Housing production is a chain rather than a permit count. It begins with land whose rights are clear and whose services are feasible, then design, finance, approvals, construction, completion, utility connection and occupation. Self-built incremental housing may follow a different path from a large developer project, and infill differs from edge expansion. The World Bank urban land and housing assessment toolkit calls for describing the steps and constraints of each production type instead of naming one obstacle for the entire market. Draw at least two paths before recommending reform, such as an infill building and a peripheral development. At each step record the responsible actor, elapsed time, cost, failure risk and whether the step repeats. Not every wait is waste: some checks protect safety, the environment or neighboring rights. Advisory judgment aims to remove unnecessary delay while preserving the legitimate purpose of review, not simply to abolish oversight to accelerate a count.
Triangulate a suspected constraint with administrative timestamps, interviews across different producer types, and inspection of completed and stalled cases. One interview claiming permits take a year is not a city average. If the recorded mean is six months, inspect its spread, rejected applications and resubmissions; a mean can conceal a group that takes two weeks and another that takes two years. Measure elapsed time between explicit start and end points, and separate agency decision time from applicant time needed to complete documents. Next draw a critical path: which stage, if shortened, would actually bring delivery forward? Some approvals can run in parallel, so adding every stage's duration may overstate total elapsed time. A bottleneck need not be temporal. Serviced land, finance capacity, materials or contractors may be binding. Faster permits do not automatically mean more completed homes if another constraint remains.
Rank constraints with an explicit matrix rather than the loudest complaint. Give each hypothesis three separate descriptions: possible effect on habitable units, strength of evidence, and the authority's ability to change it within the decision horizon. Do not multiply ordinal labels into a precise-looking number; use them to prioritize tests. A water-network extension constraint might affect every peripheral scheme yet require costly, slow investment. A repeated administrative document request might affect fewer projects but be easy to test and amend. The first calls for a long investment plan; the second may warrant a rapid procedural pilot. Ask who bears cost and receives benefit. Waiving a developer fee may raise a developer's margin rather than lower housing prices if supply remains constrained in a scarce location. When a planning or finance rule changes, test effects on rental supply, refurbishment and incremental building rather than only new homes for sale.
Consider a numerical example: across 20 completed schemes, the median approval wait is 90 days, median construction time is 300 days, and median service connection is 45 days. Do not add the three medians and declare a typical project takes 435 days. The medians may come from different projects, and connection work may begin before construction ends. Recover each project's dates, calculate time from complete application to occupation, then examine stage contributions along each path. If only five schemes on unserviced land incur an extra 240 days, the policy question is not 'how do we cut every permit average?' but 'is development on this land preferable to a serviced infill option?' The answer could change if serviced infill land is costly or edge infrastructure can be financed without displacing households. Data then support a concrete alternatives comparison rather than automatic blame of one agency.
Separate a pilot reform from a causal claim. If a one-stop approval desk is followed by shorter waits among its users, easier files may have been selected into the pilot. Establish a baseline for comparable applications, compare similar projects, or roll the process out in stages if feasible and fair. Monitor final delivery time rather than approval alone, along with objections, safety and service costs. Record the risk of accelerating construction on hazardous or unserviced land and set a stop condition if safety worsens or later delays grow. State the boundary of the advice: interpretation of local permit law, changes to law and formal approval decisions belong to authorized bodies. An adviser can describe the problem, evidence and alternatives, but does not issue a permit or legal opinion simply by preparing a constraints matrix.
The deliverable is a one-page chain map: a row for each production path and stage dates, a time range, and an evidence source. Below it list the three constraints worth testing first, the data custodian for each, a reversible small trial, and a success measure linked to occupied homes rather than permits alone. Give each constraint a competing explanation. The team may suspect approval time causes scarcity while post-approval construction finance is the binding problem. State a switching rule: if approval waits fall but starts and completions do not rise after a reasonable lag, revisit the permit hypothesis rather than declaring final success. If completions rise but the target group's housing costs worsen, aggregate supply success is insufficient; reassess access and affordability. Such conditional review makes the recommendation capable of learning and correction.
Lesson summary
Distinguish observed delay from its cause and rank supply constraints by time, impact and tractability.