Because the incentives are set up such that it is more rational for governments to have a predilection for predictability over risk - even if the predictability means it will take much longer and cost much more.
That, I think, is the key observation that the article essentially ignores in favor of trivial sidebar issues (like acronyms for procurement systems that haven't been used in three years). Governments are structured as large, top-down bureaucratic entities; their projects, unsurprisingly, tend to be structured as large, top-down bureaucratic endeavors. Some of this is due to structural concerns (fund-based accounting imposes limitations on how you procure and encumber expenditures) and some due to substantive concerns (agencies do not have the space to "pivot" their operations; when you're responsible for public safety and welfare, incremental iteration can be a pretty risky play); in either case, it certainly, as you point out, biases governmental organizations to caution, trading money and time for predictability.
Unfortunately, as we know, in project management unameliorated risk is a compounding factor, and iteration and hypothesis testing are some of the few risk sinks we know about in practice. Squaring that particular circle is the holy grail for practitioners inside the government and out.
Unfortunately, as we know, in project management unameliorated risk is a compounding factor, and iteration and hypothesis testing are some of the few risk sinks we know about in practice. Squaring that particular circle is the holy grail for practitioners inside the government and out.