Confessions of the Deep State — Part Seven: The False Promises of Centralization

Welcome to the last entry in this series! If this is your first time reading one of these articles, these are my personal confessions of problems that we really need to solve in the federal government. Sometimes I even suggest solutions. I like to think that this stands in contrast to the various levels of hyperbole you might hear from news outlets and radio commentators, most of whom are frankly just appealing to the lowest common denominator. And if you’ve just arrived at this series, here’s Part 1, Part 2, Part 3, Part 4, Part 5, and Part 6. If you like what you read here, please check ’em out.

Oh, and in case you missed it, there is no Deep State as far as I’m aware. At least not in the shadowy cabal sense. Sorry. There are a large number of intersecting interest groups in and out of government, but they are neither centrally organized nor integrated to form any kind of organization that fits the “Deep State” motif so beloved by Mr. Trump and his allies. There, that’s one confession down — now on to the real meat of the article.

There is a serious problem with the concept of centralization in the modern US federal government. I’m not talking about centralization of authority — that’s common to most forms of government and is in fact the whole point of government in the first place. No, I’m talking about the next level down, where government operations get totally screwed up thanks to the implementation of business ideals of centralization. Specifically, the constant nagging idolatry of shared services.

Let me set the stage a bit. As you may or may not know, each government agency has a certain number of people in authority positions who are political appointees. This means they change with each administration (and usually several times within an administration), and one of their main jobs is to promote the agenda of the President. Because they are appointed positions, they are often staffed by people who are known to the President or the President’s party, usually in return for party loyalty or to pay back a favor, and sometimes even because the Executive Branch needs some new expertise. And as a result, these folks quite often come from private industry.

These appointees usually come into the government with at least some idea of how a large organization works — they are surprisingly infrequently just political hacks. Mostly they’re private industry leaders looking to pad their resume, or former high-ranking employees of county/state governments. Unfortunately, their idea of how things work doesn’t always translate to how the federal government works, and in no area is this more apparent than in the fascination these people have with shared services.


A shared service is quite simply when one office adopts a responsibility on behalf of several other offices. Human Resources is a classic example: many companies do not have a separate HR office for every team or division. They might have a local representative, but usually that person reports to corporate HR. The corporate HR office is responsible for maintaining all the personnel records, processing all the paperwork, getting all the insurance in order, etc. They are performing a shared service on behalf of all the satellite offices, rather than have each and every smaller office maintain a full, dedicated HR team.

This is not at all the way things often work in the federal government. A Cabinet-level agency or department might have multiple smaller agencies or departments under it, and each of those agencies probably has their own HR group, IT group, procurement team, etc. The very first thing a great many politically-appointed government executives think when they arrive is, “What the hell, why am I paying for sixteen different HR offices?!”

The second thing that a great many government executives think is, “Why doesn’t this work like [Insert Previous Job Experience], where we only had one HR office?”

The third thing these executives think is usually, “It must be because government is incompetently inefficient. I’ll fix that, by gum! We’re implementing shared services!” And thus begins another round of attempts at making government more efficient, perpetrated by people who don’t know how government works.

They normally fail, if your definition of “efficient” is:

  1. A shared service that costs less, and/or
  2. A shared service that provides equal or greater service than agencies had previously received.


Understanding why this fails, when it quite often at least nominally works in the private sector, is one of the keys to understanding the challenges of the federal government. If you’ve read the other entries in this series, you might already be able to piece together the answer. For those of you who don’t want to invest an hour of your time, here’s a quick overview of three points that normally doom the shared service approach before it really gets started:

  • Government agencies cannot hire the ideal number of staff. See Part 1 of these Confessions for more details.
  • Most agencies have intricate, internal bureaucracies that do not lend themselves well to standardization. See Part 3 for more details.
  • Government agencies have hard funding caps and issues with being efficiently funded just in general. See Part 4 for more details.

This adds up to a situation where agencies cannot easily hire the staff necessary to support shared services, cannot easily acquire the funding necessary to support shared services, and may not be able to provide shared services to as many groups as a new executive might initially assume. Add to that one final point:

  • Some agencies answer to no one but the President and Congress.

Let’s walk through this issue with an example — let’s say a Department tries to centralize their procurement offices. Instead of one procurement team per agency, they try move everyone to one procurement shop for the whole Department.

This seems like an easy and natural fix, right? Procurement is a labor-intensive job in the federal government; a LOT of effort goes into issuing contracts and buying things, because of all the paperwork involved. It seems like it would be a good move to consolidate the various procurement teams, manage the group centrally, and be able to pool effort and ensure that the Department isn’t issuing duplicative contracts. This should, in theory, reduce the volume of procurement actions and make the procurement process more efficient.

Now, what would actually happen in this situation? Let’s run down the necessary actions and consequences in sequence.

  • One of the agencies, probably the Office of the Secretary or a similar central office, has to host the centralized shared service.
  • The hosting agency has to acquire the staff to support the shared service. They can do this by either hiring contractors or hiring more federal personnel. Additionally, they cannot hire contractors for all positions, due to conflicts of interest, so some additional federal personnel are required in any case.
  • The hosting agency must then ask Congress to increase its personnel cap for federal workers in order to accommodate the larger workload. If the hosting agency cannot increase their federal cap, they must hire contractors. This will increase the cost of service by between 1.5x and 2x because contractors are more expensive than federal personnel.
  • The hosting agency must therefore increase its budget to pay for the shared service. This typically means “taxing” the agencies who are using the shared service to pay for expenses and overheard.
  • This means that, even if large numbers of contractors are not employed, the agencies are now paying more for the service than they were previously, since before they just had to pay for the bodies, not the overhead tax. This immediately eliminates one aspect of efficiency.
  • The agencies now have excess personnel slots. They have no incentives to reduce their personnel cap or reduce their funding requests (see Part 4 of this series). They will therefore keep the personnel slots and increase their funding request (unless prevented from doing so, which rarely happens). This eliminates another aspect of efficiency, but oddly enough meets the mail for one selling point of the shared service — freeing agency resources to do other tasks.
  • Meanwhile, by moving services to a centralized location (or worse yet, off-site), and/or replacing federal workers with external contractors, the shared service almost always decreases customer satisfaction. It’s very easy to go yell at Jim in IT down the hall if your new computer doesn’t work. It’s much harder to beat your way through a phone tree to talk to a customer service representative for an IT shared service for the same problem.*
  • Shared services are also commonly less responsive to individual customer requests — their budgets are set centrally, meaning that local agencies aren’t able to boost spending for things they specifically need. So getting a shared IT service to adopt a new technology is going to be a much harder sell than trying something new in a local IT team. Requests start going into queues instead of going to project teams.
  • Finally, because client agencies are chipping in to the combined shared services budget, and are typically doing so at a set rate, the return on investment for the service rapidly becomes opaque — it’s very hard to see what your money is actually buying you, as opposed to what your money is buying everyone, in most shared service models.

The end result is usually a situation where shared services actually increase the amount of money everyone is spending (in absolute terms) while decreasing the quality of service. This breaks both sides of the argument for moving to shared services, and yet this is one of the most common managerial changes that new federal executives try to implement.

Add to this an additional major impediment — federal appropriations law. Did you know that not all funding in the federal government is equal? That some pots of money come with expiration dates, while others are available forever? Or that you legally can’t just consolidate funds into a central pot in order to dole out money saved via efficiencies? If you didn’t know that, you’re not alone, because most political appointees don’t either.

Long-term federal staffers know all of this, which leads to a more insidious problem: foot dragging. Put simply, instead of telling the boss “no,” which rarely works, most agency managers will simply say “yes, and,” then proceed to list all the challenges that will need to be overcome before a shared service solution can be put in place. This typically results in a management quagmire that occupies months or years of executive attention before the boss either forces a solution through or gives up on the whole thing. Net result: a very long period of wasted political capital, resources, and working hours.

And finally, sometimes agencies are just not accountable to anyone. The FAA is like this, for example; it’s a DOT agency in name, but a totally independent entity in practice. The DOT can set policy if it wants to, and FAA can happily tell them to go suck an egg. In such a situation, shared services have a lot of trouble because they depend on all of the agencies chipping in. Having a golden calf hanging out on the fringes tends not to help.

So in summary, shared service implementation in the federal government usually result in:

  1. Higher costs for everyone.
  2. Worse service for everyone.
  3. A lot of wasted time and resources while everyone is trying to figure out how to implement the new service.**

This is hardly ideal. It’s even less ideal if you’re dealing with tens of thousands of people and billions of dollars of public money.

From the management perspective, the core problem is that people with experience in the private sector seem to be dumped into the public sector with little to no training in the organizational issues that come along with public sector work. I can’t count the number of times I’ve seen new federal managers shocked by the length of the federal hiring process, the travails of dealing with unions, the legal minefields of contracting, or the hoop-jumping exercise of federal accounts management. And so they flounder around, fight the last war, and push policies that they think ought to work… but don’t.


The mania for shared services is largely the result of bringing people in who want to run the federal government “like a business.” This is one of those sound bytes that gets endless replays and resonates with a large percentage of the citizenry. It’s also totally stupid.

The federal government is not a business, on any kind of fundamental level. It does not have the same goals, it doesn’t have the same funding strategies, it doesn’t have the same internal structures, it has way more legalese, and it is not malleable. To change things in government literally takes an act of Congress, an invested President, or a highly empowered Department head — or, for preference, all three. These circumstances rarely occur. In any instance where they do not occur, trying to run the government “like a business” will fail.

That being said, these private industry wonks are not wrong in one way — a properly run shared services group could conceivably save costs and/or increase efficiencies in most federal departments. It’s just impossible to do this right now because of the way the government is currently organized. So we have to change that if we want to gain efficiencies from shared services. Here are five possible solutions to this problem:

Solution #1 is to train executives. You know how everyone complains that there should be a test before you have children, or become an adult? Same concept. We should not expect private sector leaders and experts to instantly understand government red tape. That’s not fair. A two week course or something similar could go a long way towards ensuring that incoming leaders understand at least a little bit about how a federal agency functions — and how it doesn’t. This will save a lot of frustration later on, and help new executives make informed decisions instead of trying to run down the failed paths of their predecessors.

Solution #2 is to remediate the way the Cabinet-level departments manage personnel. Right now it’s very difficult for Cabinet-level departments to move people around their sub-agencies. This is because, as previously mentioned, agencies have personnel caps. I understand why Congress is concerned about personnel caps, but here’s a wacky idea — how about we raise them up a level? Instead of capping personnel at the lower-level agencies, cap them for the Cabinet-level department. And then let the Department decide how to allocate personnel within its child agencies. That would wipe out a huge portion of the problem of personnel allocation and re-balancing.

Solution #3 is to assign local representatives of the shared service group to each client agency. This seems simple, but it’s often overlooked and kind of difficult due to the way agencies manage their cube farms (seriously). This would solve some of the communications delay problems and give local service users a convenient target… um… I mean, of course, a convenient point of contact that does not require phone calls or emails.

Solution #4 is to ensure that client agencies only pay costs that are representative of their service usage. Averaging costs throughout the agency always increases costs across the board. Keep costs specific and localized instead. This takes more accounting hours, but it’s worth it for the agencies and ensures that people know exactly what they’re paying for. It’s also much easier for a shared services administrator to judge how well they’re performing if they have this data for comparison against the agency-level efforts. Granted, this is probably the exact reason why so few shared service models seem to do this; no one wants to be held accountable.

Finally, Solution #5 is a radical idea — don’t bother with some shared services at all. Simply set guidance standards for agencies where necessary, police them with audits to ensure adherence to federal policy, and let them run their own little operations. This doesn’t work as well for things like cyber security, but for procurement and contracting it works beautifully. It will also save time and resources in those areas where agencies cannot implement shared services — grant-making, regulatory oversight, things like that where legislation prevents a shared service model.


This concludes the Deep State series. If you made it this far, you’re a trooper and my hat is off to you. A lot of this stuff really epitomizes one of my favorite life lessons — boring things are sometimes extremely important. None of these discussion have been sexy. Not a single explosion was mentioned, or even alluded to in most cases. No car chases occurred, and there wasn’t a single discussion of torrid passion or saucy scandal. But nonetheless, these have been important subjects. These are problems that the federal bureaucracy needs to solve, at least if we want to maintain a functional state.

It’s worth noting that this is exactly the opposite of the goals of many, many politicians right now. Some people want to “starve the beast.”

Don’t let the TEA Partiers and Freedom Caucus fool you, a workable bureaucracy is essential to the American way of life — if only because there will always be bureaucracy. The b-word is just what you naturally get whenever you try to organize more than a couple hundred people; it’s as inevitable as sunrise. And yes, it’s always going to suck just a little bit. But the functionality of a nation that works vs. one that doesn’t lies between the bureaucracy being a minor inconvenience and the bureaucracy being an actual obstacle… and we are in danger of moving towards being an obstacle.

Wish us luck, folks. We need it.

*There are actually a bunch of other reasons why service quality almost always decreases — new Service Level Agreements that are more generous to the service provider, decreased attention to individual customers, lack of personal interaction, less involved management, fewer incentives, etc… All of this should be a serious concern for anyone considering a shared service path.

**I say “wasted” because if the service was cheaper and better, this would be a good use of time and resources.

***Because they’re mostly getting by on bailing wire and duct tape already. Seriously, Americans have absolutely no idea how precarious the resource situation is at a lot of agencies.

Searching for truth in a world focused on belief.