Our technological systems are becoming more complicated. In some cases, even so complicated that the experts involved in their construction don't fully understand them any more. But too often, when we think about technology, we focus on certain kinds of systems: computers and large machines. But there are many other anthropic systems that might be considered technologies. And this includes our laws. Essentially, our legal codes are complicated technologies, growing and becoming more interconnected over time.

But how complicated are our laws? In a working paper titled “Measuring the Complexity of the Law: The United States Code”, Daniel Katz and Michael Bommarito of Michigan State University recently set out to measure exactly that. They attempted to quantitatively measure the complexity of the United States Code, using what is roughly a metric for how hard it is to understand it. The U.S. Code is essentially the collection of all federal laws, and consists of 51 Titles, or sections, that each deal with different topics. For example, Title 11 is related to bankruptcy, Title 26 is our tax code, and Title 39 deals with our postal service. Since there are many sections with different topics and styles, comparing the complexity of different Titles is a natural means of examining differential legal complexity. But how to do so?

While there are many ways to measure complexity, Katz and Bommarito focused on rule search and rule assimilation which are, respectively, “How complex is the task of determining the rule or set of rules applicable to the conduct in question?” and “How complex is the process of assimilating the information content of a body of legal rules?” While there are many other ways to measure legal complexity, such as the cost of compliance with laws, they chose to limit their inquiry to just how hard it is to understand this legal code.

But even by only examining this type of complexity, there are many ways to examine this. For example, the more interconnected a section of the law is, the harder it is to truly assimilate and understand the law's implications. Here is a snapshot of a particularly highly interconnected section, a short part of the United States Code that deals with the Alaskan Railroad:

(b) Federal surplus property disposal; withdrawal or reservation of land for use of Alaska Railroad
The enactment of this chapter, actions taken during the transition period as provided in section 1204 of this title, and transfer of the rail properties of the Alaska Railroad under authority of this chapter shall be deemed not to be the disposal of Federal surplus property under sections 541 to 555 of title 40 or the Act of October 3, 1944, popularly referred to as the “Surplus Property Act of 1944” (50 App. U.S.C. 1622). Such events shall not constitute or cause the revocation of any prior withdrawal or reservation of land for the use of the Alaska Railroad under the Act of March 12, 1914 (43 U.S.C. 975 et seq.), the Alaska Statehood Act (note preceding 48 U.S.C. 21), the Alaska Native Claims Settlement Act (43 U.S.C. 1601 et seq.), the Act of January 2, 1976 (Public Law 94–204; 89 Stat. 1145), the Alaska National Interest Lands Conservation Act (Public Law 96–487; 94 Stat. 2371), and the general land and land management laws of the United States.

As can be seen, it is highly connected to other parts of the code, increasing the complexity of understanding it. So one way looking at complexity is by seeing how many connections a section has to other sections.

And they looked at a whole variety of other metrics as well. They examined the hierarchical depth of the different Titles (section, subsection, etc.), in order to see how people might navigate its structure; the number of words in different Titles; and even the average word size.

So which were the most complex sections of the U.S. Code? According to their rankings, overall it seems that Title 42 (Public Health and Welfare) is the most complex and interconnected, as is the Internal Revenue Code, or Title 26. Which shouldn't be surprising to anyone who has tried to do their taxes.

A fascinating next step would be to see the change in complexity over time of different Titles, as well as of the entire U.S. Code. But even without that, it is incredible to see how much variation in complexity there is across our federal laws. I recommend checking out the full paper; it is a great read.