Confessions of the Deep State — Part Three: Death by Red Tape

Good day, dear reader! If you’ve just discovered this little article, it’s the third of a series of essays on some of the actual, serious problems I’ve encountered as a member of the federal workforce. Also, full disclosure, as we discussed briefly in Part 1, there is no “Deep State” in the sense that the alt-right would like you to believe. My apologies for luring you into a (hopefully informative) discussion under false pretenses, but it was such an easy title.

In this edition of the Confessions, we’re going to be talking a bit about one of the most familiar icons of Big Government — red tape. For any readers who are new to the English language and/or American idioms, or who have spent a lifetime under a really worry-free rock, “red tape” refers to the endless strata of bureaucratic forms, policies, and procedures you have dig through to get anything done in dealings with the federal government. They often have names like “Form 1037-A (Part 2 [Revised] of 620)” or “SMEGLE Version 92” or, if you’re really unlucky, “NIST 800–53.” They typically require hours of close study with a thesaurus, a lawyer, and a signed contract with the devil before they start making any sense. And they are utterly endemic in federal life.

Today we’re going to talk about how they come about, why they continue to exist, and what can be done about them. Buck up, it’s not all bad news. It does, however, require a brief diversion into a discussion of the structure of government organizations, and more generally an answer to the eternal question: What the actual hell is bureaucracy good for, anyway?

This requires a bit of explanation. To start with, you have to answer the question, “what is it that government is supposed to do?” And the historical answer is: things that have to get done. Government usually gets involved when a social need exists for X and the private market does not provide X to the people who need it at what we, as a nation, consider a reasonable price, quantity, or level of availability. There is some definitional wiggling you can do, but that’s it in a nutshell.

Why laws? Because we’re not all angels, and every human society in history has required a central, legitimized authority to stop people from taking advantage of each other.

Why a national army? Because mercenaries are too expensive and unreliable.

Why national safety standards? Because we can’t trust business entities to judge themselves, and because business entities love to have authoritative targets they can claim to have hit.

Why subsidized healthcare? Because eventually everyone gets sick, which breaks the normal market models, and also because poor people can’t participate in the market and we think that shouldn’t be a death-penalty offense.

And so on, and so forth. Government exists, in fact, because we think it ought to at a societal level. Otherwise the whole thing comes tottering down in a few years of rebellion. But while it exists, democratic government tends to do things that most of us agree are necessary, usually because if they don’t get done, someone dies.

This is the core difference between a business and a government bureaucracy. Businesses exist to provide services at a profit. If a business fails, it affects the owner, employees, and shareholders; everyone else just gets their stuff from someone else. When a government agency fails, someone usually dies. Therefore US government agencies tend to adopt a “must not fail” approach to operations, as opposed to a “must turn a profit” or “must not lose money” approach, because despite everything Ayn Rand tried to do most people still tend to think that letting fellow citizens die is a Bad Thing. After all, the next head on the chopping block of life might be yours.

Add to this the concept of bureaucracy. A bureaucracy is any large organization that has grown to the point that it requires internal rules and processes to manage its day-to-day functions. Functionally speaking, this happens any time you get more than a few dozen people in a group and ask them to work together. It is a natural outgrowth of human social dynamics. As such, a sufficiently large church is a bureaucracy, your local Boy Scout troop is an arm of a bureaucracy, virtually all major companies are bureaucracies, and the government is basically the great-granddaddy of all bureaucracies.

Bear in mind that having a “must not fail” approach to providing services as a bureaucracy has consequences — particularly in an environment where you know it will be hard to attract the best and brightest (because government pay doesn’t beat industry pay for the best of the best) and where you know you won’t be able to get rid of all of your dull bulbs (because getting fired from the government basically requires an act of treason). What do you do when operating under these constraints? You focus on hardening your business processes.

The Army has known for a long time that training and conditioning can take the place of intelligence in certain circumstances. You don’t have to have an entire army of Alexanders if you can train each of your soldiers in precisely what to do under certain scenarios; you just have to hope like hell you’ve thought of all the scenarios! Most bureaucratic agencies have it much easier — they invent the scenarios in the first place, then train employees to perform as cogs in the larger machine. They then try to maintain staffing levels sufficient to have at least some redundancy, so that even if one cog leaves the agency, gets a promotion, or wins the lottery and tells their boss to go suck an egg, another cog will always be available to pick up the work.

This is not an approach that rewards original thinkers, geniuses, or people who like to work outside the box (although, paradoxically, all of those people tend to do well in government service). And having redundancy in your operations means, by default, accepting a certain level of inefficiency. This means that government agencies tend to be large, with highly specialized processes intentionally designed to crush through their assigned workload given sufficient time and money. It is a brute-force approach to problem solving that does not rely on high-flyers or sudden moments of inspiration (although both of these phenomena do occur), but rather relies on working through each process step by step, in order, by the book, every time, to achieve a desired result.

A bureaucracy designed in this way requires process models that are standardized and comprehensible within the bounds of the bureaucracy. This directly results in forms, and moreover forms that have a large number of numbers in the title (because eventually you run out of words to call things, but numbers keep going). It doesn’t matter as much that they don’t make sense to the average citizen; the average citizen isn’t the target.

It’s Mildred in Accounting, who has a room-temperature IQ but who’s looked at the same acquisitions form in more or less the same format for her entire career, to the point that she can do her job by rote, accurately, and to a defined deadline.

It’s Fred in Grants Management, who has no idea what a grantee is actually planning to do with the money, but who is perfectly capable of working within his process box to make sure the numbers in sections 10, 13, and 187 of the required forms line up properly and cross-reference with the grant profile in the right way.

It’s Javier in HR, who doesn’t understand the legal philosophy behind federal diversity requirements (and wouldn’t if you explained it to him with sock puppets with narration by Mister Rogers and Morgan Freeman), but who is perfectly capable of checking resumes to find a blind veteran black female candidate to meet the hiring quota.

It is, in fact, to allow the organization to get on with the minutia of getting things done irrespective of the talents of the actual employees. In the better agencies this also enables people higher on the org chart to look at all the data and make intelligent decisions. Every bureaucracy feeds the top layers of management by grinding out a jillion form-based processes lower down and then synthesizing the information. In really good agencies those top layers have avoided the Peter Principle.

That’s the ideal world. In real life, more insidious things tend to happen.

For starters, process models tend to change on an annual basis, as Congress or agency planners dream up operational requirements, and as new management comes in to make its mark. This leads to requirements for new processes and forms. But the old requirements rarely fade entirely away, and the old processes and forms tend to stick around. This leads to a slowly accumulating pile of paperwork that has to be done before you can get anything else done. Eventually the old stuff tends to get phased out, but this rarely happens due to deliberate action; it’s usually more of a situation where people just stop following one set of instructions by unspoken agreement.

For example, there’s a federal process called Information Technology Capital Planning and Investment Coordination. It exists to track what agencies are doing with their IT funding, both to increase the public visibility and to allow Congress to make funding decisions, and some of the data it collects are publicly available here. The reporting forms used by this process are myriad, Byzantine, and sometime particular to individual agencies. As a result, the amount of time required to fill out all the reporting forms tends to increase every year, and only very rarely ever decreases — it takes an enormous effort to figure out what to phase out, and it’s always easier to just add in new fields to capture the data you want to look at this year. But the old forms hardly ever go away, even if no one looks at the data they collect anymore.

Accordingly, these reporting activities have evolved from being a part-time exercise to being a full-time job practiced by people with a rather arcane knowledge of all of the processes and procedures required to get through the annual forms… and this process is required by law. There’s no getting around it. So instead of the process fulfilling its stated goal of collecting IT information, it is increasingly serving primarily to increase the federal or contractor workforce requirements of IT offices.

This leads us to another issue: you quite often run into the problem of people mistaking the process for the end result. As a particularly notable example, government cyber-security is almost laughably bad in many (if not most) agencies. People think government cyber-warriors are working out of LED-lit concrete offices with unusual glass furnishings and huge wall-mounted monitor arrays, surrounded by brilliant former black-hat hackers who are just trying to make things right with their adorable daughters. And sure, there are a couple of offices like that. But most cyber-security folks work on old PCs, try to manage hodge-podge networks, and devote most of their time to filling out the endless audits, approval forms, system reviews, privacy impact statements, and security scan appraisals that are required by law of all government agencies to meet federal cyber-security regulations.

This doesn’t start to get seriously scary until you realize that:

(A) Congressmen on the whole are not computer experts, and the laws they pass to regulate cyber-security might not be totally up to date, and,

(B) no one is actually doing any cyber-security because they’re too busy filling out forms.

As a direct result, most federal agencies’ computer networks are penetrated on a daily basis by the Russians, the Chinese, and the occasional script-kiddie who got bored enough to read the publicly-available federal cyber-security guidelines and made all the right deductions. The cyber guys can’t stop them, because it is increasingly not their job to do so. It’s their job to fill out the forms. That’s just an example, but it’s a very common case of mistaking the cart for the horse.

Last but not least, a bureaucracy based on process models acquires an enormous burden of inertia, such that it becomes incredibly difficult to change how an agency does business. This makes sense if you remember one of the key points of the Iron Law of Institutions: the members of an organization will always be primarily concerned with making sure their own positions are secure. Remember Mildred from Accounting? How do you think Mildred is likely to feel about a form change at this point in her career? She’s been doing the same thing every day since Noah got off the Ark. She’s not going to want to change. Now scale up: how do you think Mildred’s management feels about having to design a whole new process for all their Mildreds every time a new law gets passed?

Sometimes, shit flows uphill. Federal agencies have almost as much lobbying power as some K-Street firms; they’re generally trusted, have an easy mouthpiece in the form of news conferences, and have immediate access to Congress. And they don’t like change; it’s expensive, disruptive, and introduces risk. When you combine these factors, it becomes easy most of the time for agencies to slow-walk radical changes. This is both a blessing and a curse. When the country is trying to do something generally palatable, it can be a curse. When you get a rank amateur in the White House and a whole host of maliciously-optimistic fools in Congress, it can be a blessing. But regardless it means that the federal government is slow to adopt new ideas.

All of this, of course, has an impact on John Q. Public. He looks at the government and sees this massive, imposing edifice of stern-faced drones behind towering bulwarks of jargon-filled forms — and it gets in the way of getting things done. What, oh what, can be done to fix this?

Well, I’m glad you asked, John! In the first place, we have to address the elephant in the room: Some of this crap is necessary.

President Trump held a news conference at the Department of Transportation not long ago where he slammed a big binder full of regulations down on a table and pointed at it as an example of bureaucracy gone bad. What was in the binder? Good question! It was the ecological impact assessment of a new highway. Anyone who likes the concept of clean drinking water is probably in favor of that assessment getting done, if they think about it for a moment. Is it likely that the assessment is inefficient? Sure. Is it likely that it will slow down the construction? It sure is. Is it probably necessary that we do it anyway? You betcha.

There is an old maxim: you can have government that is efficient, or you can have government that is accountable, but you can’t have both. A great many bits of government paperwork exist primarily to document precisely what the government is doing, just in case John Q. decides he wants to know. As a democracy, we should probably be in favor of this as an ideal practice. After all, how else is John going to know what’s going on, and thereby decide where to cast his vote? Should we just believe whoever is in office when they tell us what’s happening? Have you seen the occupant of the White House currently? So yes, I’m a little leery of any proposal that tries to fix the red tape burden by limiting government accountability.

What we can and should do is decrease the burden of internal process models at federal agencies, both for the public and for federal workers. Automation is our friend here; many government services are still run on paper, and computers can do the job MUCH faster. Efforts to make inter/intra-office approval processes more efficient would probably also be advisable. Similarly, it couldn’t hurt to do a plain-language assessment of federal forms every year or two to cut out some of the jargon and a lot of the verbiage.

Most importantly, much like a broken clock is right twice a day, President Trump did have a grip on a key problem when he asked agencies to deprecate old regulations. The way he went about it was debatable, but the idea was at least partially in the right place. Just clearing out the dead wood would make many federal employees’ lives easier and their work more efficient. The catch is that you, John Q., need to make really sure you trust whoever is in charge of burning off that brush — otherwise you’ll end up burning a live tree by mistake (like the EPA).

Last but by no means least, we should be making a special effort to identify and correct areas where the process has become a goal in and of itself. There are many, many circumstances similar to the cyber-security example example from earlier, and they all eat up valuable time and resources while failing to actually produce the correct result. Government is littered with examples of this sort of phenomenon; other nasty ones include the federal procurement process and a great many Office of Management and Budget reporting requests.

To paraphrase the late great Terry Pratchett, any initiative designed to increase efficiency will first generate more inefficiency. This will certainly be the case with trying to reform federal red tape, since any reformer will inevitably end up filling out red tape forms about the red tape they’re supposed to be removing. After that comes the process model audits, planning sessions, working groups, and management reviews — all the stuff that, in fact, looks an awful lot like bureaucratic busywork from the outside. But bureaucracy isn’t going away; the best we can do is try to make it rational again.


Dear reader, I hope this was at least a little informative and gave you a better view into the trials of federal agencies. If you’re interested in the other stories in this series, click the links for Part 1 and Part 2.

Searching for truth in a world focused on belief.