Summary:
Whether or not its cheaper to build a custom home or buy a newly constructed home is a complicated question with many variables. We take a look at both the qualitative theory and the raw data to compare three different contracting options to help inform the decision making process for prospective homeowners. We find that substantial savings can be made by building a custom home instead of buying a newly constructed home, but certain risks have to be carefully controlled and mitigated.
Qualitative Analysis:
New construction can broadly be divided into three categories:
Each of the main three types of companies involved takes a profit margin that is directly related to the amount of risk they are undertaking. Developers tend to get the largest profit margin because large scale land development can be risky since it often involves securing zoning variances or land subdivision which are not always guaranteed to work out as planned. General contractors tend to get a medium sized profit margin because there's a lot of uncertainty in site conditions, code enforcement and material costs. Finally, subcontractors tend to get the lowest profit margins because there's minimal uncertainty in materials or labor costs.
Category A - Houses Built for Sale (by Developer):

Category B - Custom Built Houses (by general contractor):

In this scenario, the owner takes on the role of the developer and eliminates their cut of the profits. Note though, that the owner must also take on the risk of the land being undevelopable.
Category C - Custom Built Houses (by owner-builder):

Finally we have a scenario where the owner takes on both the role (and risks) of the developer and the general contractor.
The Cost Savings:
So in theory, a homeowner can save quite a bit of money by cutting the developer out of the equation and assembling the design team, financing and general contractor his/herself. What does that look like in terms of savings?
We looked at an example where a house is built to be worth $500,000 when complete. As you can see , it's possible (in theory) to save quite a bit of money by eliminating the developer ($125,000) or the General Contractor ($75,000).
The Risks:
Building your own house does not always pan out this way. Both land development and general contracting are highly skilled professions that take time to develop and perfect.
The key aspect to understand is that most of the value that developers bring to the table is based on their pre-existing relationships. They have key working relationships with people in financing, design, and local government that allow them to reduce the friction in the home building process and get things done efficiently. They are able to leverage these contacts to reduce their risk. For example, before buying a large parcel of land for a subdivision, they may have conversations with their civil engineer to assess feasibility, and the town official in charge of signing off on the plot subdivision to ensure that it will go through smoothly.
Controlling Land Development Risk:
For building a single family home in a rural location, land development risk is fairly minimal. That risk only increases if you need to subdivide a parcel of land or get a zoning variance for some construction that is not permitted by right. For this reason, it's vitally important that you or your real estate agent perform a comprehensive land feasibility study before placing an offer to buy land to build on.
One of the larger remaining risks for a custom home owner is the site development costs. It's important to fully understand the scope of that work and secure written bids before proceeding.
Controlling Construction Risk:
Whether you hire a general contractor or become an owner-builder and only hire subcontractors, it is extremely important to properly vet your builder during the bid phase.
Many general contractors will try to outsource their risks onto the owner, such as material costs, construction delays, and change orders. Stay tuned, for helpful tips and strategies for controlling those risks during the bid stage.