Inspired by Episode 43. THE CENTRALIZATION/DECENTRALIZATION DEBATE – THE FEDERALIST PAPERS
By Tom Galvin
Listen to Tom’s sidecast here:
Greetings, and I hope you enjoyed listening to us debate the merits of centralization and decentralization in organizations on Episode 43, where we framed the debate using as lenses Federalist Papers #9 and #10, authored by Alexander Hamilton and James Madison respectively. If you have not had the chance to listen to the Episode, I encourage you to do so.
Although we took sides and argued in teams for the purposes of the podcast, we acknowledged both benefits and risks to organizations that choose either one as a preferred way of doing business. Both can also be guiding principles. De-centralization as a guideline or strategy is thought to promote innovation and greater participation. Organizations that de-centralize as a matter of course may view centralized hierarchy as old-fashioned, complacent, even ossified. So, what about the virtues of centralization? One of the main arguments for greater centralization in organizations is that it brings greater efficiency.
But What is Efficiency?
Ask different people what efficiency means to them and you are likely to get different things. In its most elemental form, efficiency is classic Taylorism — meaning more output for the same input. Taken further, this can lead to Holy Trinity of economics: better, faster, and cheaper. People often emphasize one over the others depending on the commodity or service in question or the location. I recall how efficient the train systems were when I lived in Europe in the early 2000s. At the time, you could always count on the trains running on time or knowing precisely how late a train may be. As applied to centralization, efficiency can also be a euphemism for consistency or reliability. When one goes to a McDonald’s anywhere in the world, one expects the hamburgers to have the same quality, served speedily and at low cost compared to other dining options. You may be thinking that this sounds a lot like effectiveness. Perhaps, yes. This is because while organizational science may separate efficiency from effectiveness, economists are more likely to use them interchangeably. In my experience, this is one of several reasons why efficiency is a confusing term. I will use the economic view as the basis for my talk.
So let’s address efficiency as an organization goal. With thanks to Dmitrijs for this insight, efficiency as an organizational goal is a relatively new idea. Prior to the late 1800s, organizations were not so concerned about efficiency as they were about continuity, power, or building relationships and networks with other organizations. The onset of industrialized competition changed the nature of competition, and organizations had to produce more and quickly, or lose out in the marketplace. The Efficiency Movement, which Frederic Taylor contributed to, fostered the adoption of scientific management beyond industry and into other types of organizations.
It is important to acknowledge that embedded within the concept of efficiency are contradictory ideas. Consider government services, such as drivers licensing. Consolidation of licensing offices is often done to save money and ostensibly make things more efficient by reducing the money spent on facilities and manpower in favor of a single larger regional office with modern capability. But to the public, such moves may have the opposite effect, causing citizens to have to drive further to the nearest office to renew their licenses and wait in longer lines. In such cases, the quest for efficiency seems rather one-sided, does it not? Certainly, the citizens are asking, “efficient for whom or in what way?”
But in the end, what I’m going to show is that no matter how you define or describe efficiency, your pursuit of it leads to one thing and one thing only. Some form of centralization.
Three Common Variables of Efficiency – Time, Cost, and Quality
At its core, efficiency is a measure of how much input produces how much output. For many years in manufacturing, the conventional wisdom was: you might want things fast, cheap, and perfect but can only hope to get two out of the three. The question became, “Which two?” If you are an organization providing goods or services and you emphasize fast and cheap but your customers expect perfect, you may think you are efficient but your customers may not and this difference in perception will likely impact sales.
The trouble is that even economists have different conceptions and definitions of efficiency and thus have multiple ways to describe it. Let’s look more closely at fast, cheap, and perfect – or put another way, time, cost, and quality. As consumers, we would probably all agree that faster, cheaper, and better are preferred over slower, costlier, and poorer quality. That is not exactly challenging any theoretical boundaries, is it? But do we always seek the best possible combination of all three? Experience tells us no, that we may use these measures to make choices between the goods and services we may buy, but in truth we accept a certain degree of inefficiency in many circumstances (as an aside, we touched on this in Episode 19 on organizational learning and bounded rationality). When hungry and in a hurry, eating at a chain burger restaurant, we prioritize fast and cheap so long as the burger’s quality is “good enough“ that it does not lead to health problems later. When we have more time and are not with tired and fussy children, we may be willing to pay more for a quality burger at a fancier restaurant, the kind with cloth napkins, and will be more willing to sit and wait for it. But when we visit a steakhouse? I don’t know about you, but my steak better be served exactly as ordered. I’ll accept the higher cost, which might be the equivalent of four meals from the hamburger chain. So again, as consumers, we establish a balance of these qualities of efficiency based on personal expectations that can vary in innumerable ways.
Two Models of Efficiency – Business and Productive
Let’s now situate ourselves in the organization producing the goods or providing the services. Assume we are a widget factory, and we produce 100 widgets per hour at a cost of $50 apiece, with each widget selling on the market at $100 apiece. Like our forebears who produced scientific management and other ideas, we want our factory to be more efficient, so we might pursue making things faster (like 120 widgets per hour) or find cheaper materials or labor to cut cost of production to $40 per widget, or do something to improve quality without affecting time or cost.
One model of efficiency is called business efficiency, the straight ratio of total measurable output to total cost. This is the way efficiency is often used in the financial industry. As applied to our factory, it represents the total amount of revenue over total cost. So, business efficiency over the year would be based on the costs of producing and revenues from selling 200,000 widgets. If sales go down, you produce fewer or lower purchase prices. If sales are booming, you produce more (perhaps spend the more for a new factory), or raise the price to increase revenue. In government service, the business efficiency view is similar, but the aim is to ensure the consistent and reliable delivery of service to all citizens who need it. The government may still be concerned about cost and wants those to stay down, but they are more likely to hear complaints if a worthy citizen had undue difficulties receiving service.
A different, and perhaps opposing model is productive efficiency that focuses on benefits and costs per transaction or per unit produced. The emphasis is on time and cost, especially at the points of production. This perspective wants the assembly line to go faster, employ fewer people, or to use cheaper materials; or wants the retail store to sell the items faster while minimizing the cost of conducting the sale. [8:03 – insert this] We see this trend on-going in the US, as online sales mechanisms are seen as more efficient than costly brick and mortar stores with their limited shelf space. This makes it faster and cheaper to sell and deliver the product from the organization’s perspective. The quality of the product has nothing to do with the cost of providing it – a boutique brand widget is handled with the same productive efficiency as the cheapest one.
An important caveat with productive efficiency is what economists call the production possibilities frontier, which is the point where based on available inputs, no more of one good can be produced without reducing production of another good. That such a frontier exists is common sense. Consider a bakery producing both bread and muffins. When production is fully optimized, one cannot bake more bread without baking fewer muffins, because full optimization means that there is a resource (e.g., ingredients, mixers, ovens, whatever) that prevents increased production. Thus, productive efficiency can only achieve so much before a decision is needed about expanding the bakery.
The Triad of Inefficiency – Waste, Harm, and Overhead
But time, cost, and quality do not capture the whole picture. These are not the only variables we use to gauge the efficiency of an organization. There are also measures of inefficiency, which is not the precise opposite of efficiency. While time, cost, and quality may have some form of optimal balance, an organization can be completely and utterly inefficient in all three measures and still not be as awful as it could be. Above and beyond failure to deliver goods and services, organizations can be so inefficient that they can be destructive – to themselves, others, or the environment. Let me explain through a proposed trifecta of inefficiency variables – waste, harm, and overhead. Each have two dimensions.
The first variable of inefficiency is waste, both in input and output. Output first. Our assembly line might generate 100 widgets per hour, but if 10 of them are defective, one might not regard that line as efficient in comparison to a line that produces only 50 per hour with no mistakes. That is because 10 widgets worth of material is unusable, providing no benefit, and possibly its materials cannot be reclaimed or used.
There is also waste on the front end – the point of input. Let’s say that the assembly line, due to aging components or other factors, slows to 90 widgets per hour. Materials for production may accumulate in the warehouse. And what if that material was perishable in some way, like metal that rusts, or food ingredients that spoil. Same thing in government services, where offices may have to pay to correct mistakes and typographical errors on drivers licenses, for example.
The second variable of inefficiency is harm. I am not talking about harm brought about by low sales due to the poor quality of goods or services produced. I am referring to the second-order effects of production, of which there are two forms – harm at the point of production and harm from misuse or abuse of the product or service. The former is straightforward. If our widget factory is irresponsible and willfully pollutes the environment or injures its workers in order to sustain its 100 widget per hour rate, well guess what? Among litigation, clean-up costs, sullied reputation, and medical care for workers, the costs to the company skyrocket. The benefits of making 100 widgets is practically zero because modern consumers care as much about ethical practices as they do the product itself. Few people would buy from a company that shows reckless disregard to the environment. Likewise, governments risk litigation and censure if citizens are denied access to service voters think they are entitled to. Lawsuits can cripple local government budgets, but anything that causes significant inconvenience such as impeding access to services (like having to drive longer and wait longer in lines at the driver license center) can have repercussions, such as elected officials being voted out of office in democracies.
One may not think of misuse of a good as affecting efficiency, but indeed it does if the good’s benefit is reduced because of it. In essence, the good or service is viewed by some as inherently harmful. Consider debates in the US about firearms stemming from high-profile mass shootings. Social and political pressures aim to eliminate or restrict sales of such weapons, and counterpressures from other organizations seek to preserve their availability for their intended use. Or, the removal of the allergy medicine pseudoephedrine from drug store shelves because of it can be used in the illegal manufacture of methamphetamine. The drug was renowned for its effectiveness and safety for its intended purpose, but its manufacturers became victims of collateral damage from a completely unrelated problem. The controls placed on the drug meant that manufacturers had to replace it with other, less-effective drugs.
The third variable are the indirect costs not directly related to producing the product or service. These costs are sometimes called overhead, resources devoted to anything other than direct support to the core organizational purpose of providing the good or service. This can be a very subjective measure, and it affects people differently from processes, facilities and infrastructure, or technologies. A certain amount of overhead is typically unavoidable, of course, to provide the manpower and capacity for adequate supervision of operations, innovation, and growth. It is also important when the organization must establish and demonstrate accountability, particularly if there is government regulation or other oversight involved.
This speaks to the first form of overhead — remote headquarters staff or people in high places. Layers of supervisors, and staffs such as planners, accountants, lawyers, quality inspectors, and other experts who are not part of the production process but enable it by aiding the decision making of organizational leaders. The costs of these enabling functions are far easier to measure than the benefits, which sometimes causes them to appear excessive to third parties. After all, the special expertise held by these members can be expensive, and not everyone recognizes the value of work they may be doing. While in past Talking About Organizations episodes we discuss the virtues and pitfalls of bureaucracy, the term ‘bureaucrat’ is almost always used as a pejorative in common language. Large staffs might be associated with inflexibility, or filled with people working more to protect their high salaries than to benefit the organization. They can make for easy targets when budgets get tight. But there is definite risk to cutting them if one is not careful, as the organization may lose valuable expertise and indirectly place disruption and stress on the workforce.
The second form of overhead is excess, which is even more subjective. Excess facilities and infrastructure are one form of excess, as evidenced in the world of retail where many brick and mortar stores are closing under competition from on-line shopping options. There is also the perception of excess in perks. Their unequal distribution can be sources of tension within organizations, such as between managers and workers, or between the organization and society, such as public ire expressed over balloon payments to fired executives. A common complaint follows: Why does the leader get a big office with a massive oak desk and high-back ergonomic chair? Folks on the front lines might only get folding chairs or no chair at all. Of course, this is a complex issue far beyond the scope of this sidecast, but suffice to say that rewards and incentives are certainly up for debate whenever organizations face tough times. And then there is excess in symbols associated with corporate image and identity. We see this in historic buildings, unique artifacts of an organization’s culture, or the long, close association with a particular local community. Maybe these things provided strength and meaning. Or perhaps they became symbols of inflexibility and old thinking, incurring costs without adding benefit.
A Third Model – Allocative Efficiency
And so we have six variables – time, cost, quality, waste, harm, and overhead. But the two best-known models of efficiency are only concerned with the first three. So we need a third model.
And this model is called allocative efficiency. An organization is allocatively efficient when it precisely aligns costs and outputs. In other words, the quantity of goods or services produced match the demand. The implication is that waste, harm, and overhead are minimized, because these impede the ability to match supply with demand exactly. Allocative approaches work well when the costs and outputs of production are known in advance or reliably predicted. But here’s the tricky part. Allocative efficiency is less concerned about time, cost, or quality, so long as they are as stable and predictable as possible. After all, that is how the resources are allocated, which is where the name of this efficiency model comes from. If indeed our 100 widget per hour factory can be kept running at precisely that pace with known costs and known profits for sale, then we can confidently build a budget. And that’s the key word – budget.
The public sector uses allocative efficiency as its driving model. Unlike the private sector, where currency is the measure of nearly all variables, government uses programs as its basic unit of measure. Programs combine currency with the legal authority and direction to spend. This is critical. Governments establish programs by determining the services and costs in advance so they match. Legislatures or chief executives designate precisely how much an agency must spend to provide a precise level of service. Regardless of how much things actually cost or how much benefit the citizenry needs, programs essentially predetermine the costs and outputs and wrap them in a legal boundary. We hope that the figures are right in advance, because if they are not and money must be reprogrammed, that constitutes a second legislative action. Governments typically make such transactions costly – maybe not in terms of money, but in terms of trust and confidence in the agency leaders responsible for proposing the budget in the first place.
The effect of the allocative view of efficiency is that organizations are deemed efficient if they expend all budgeted resources while avoiding perceptions that the service was not provided as mandated. This view of efficiency can lead to some unintended behaviors such as how particular local government offices are forced to monitor the usage of paper clips and staples, while large federal government programs appear to exercise behavior deemed rather wasteful. A classic example of this is the use or lose culture in which government agencies must spend whatever they were granted by the end of the fiscal year or risk their budgets being cut the next year. The logic is that the agency was granted too many resources and … obviously … can afford to provide the prescribed benefits with less. So if your agency has funds at the end of the year, you better find a creative way to spend it! This has been known to be an undesirable practice, but the allocative efficiency model driving the budgetary process makes it really tough to avoid!
I could go over many particulars but suffice to say that the complexity, uncertainty, and political aspects of government programs make it virtually impossible to allocate resources properly in advance, and the demands for government services regularly outstrip the allocated resources. But this is what makes allocative efficiency so attractive – it provides concrete measures of input and output according to a standard defined by the government. Whether that standard is acceptable to voters is another question.
A Difficult Conversation that Converges One Way
My own experience suggests that these three models of efficiency – business, productive, and allocative – bring about natural tensions in organizational decision making because they measure organizational performance in conflicting, and often highly contested, ways. There is the so-called strategic or operational perspective expressed in the business efficiency model. This involves using definitions of efficiency to make important decisions about expansion or reduction, new products and new markets, or the discontinuance of either. Meanwhile, the productive efficiency model drives leaders immediately to attempt to put a price tag on everything. Costs on the production line, hiring, paying, training, human resources, logistics and transportation, everything gets priced in some way, which favors concrete things over the less concrete (reputation, community relations, fair treatment of workers, environmental stewardship, etc.). And then there are those employing the allocative efficiency model, where decisions are rooted in zero-sum language. Doing something new means removing resources from something else, or if that is too hard, then deferring the action to a future time when the budget is constructed anew. These three perspectives represent three completely different ways of framing problems and options. What looks good from a strategic perspective within one model may be technically problematic or wasteful in the others, and so on.
Yet buried among these perspectives is a common theme. Regardless of which form of efficiency an organizational leader favors, it ultimately drives the leader down a path of centralization. It just does so in different ways – by strategies and plans, by measures or performance, or by control over and distribution of resources.
This is why during the episode, I tried to argue that the natural tendency of organizations is to centralize, even when their desires are the opposite. The different drives for efficiency view centralizing authority and oversight as more efficient than options that diffuse them. Thus, once efficiency is essentially optimized and nothing more can be done, then inefficiencies creep in to the point where the organization has little option but to de-centralize. But this is disruptive and leads to much uncertainty as leaders cede control or try to make sense of the chaos they perceive within the organization. Members of the organization become uncomfortable with the churn of change, and may gravitate back toward the familiar old ways even though they may have previously cursed them.
Efficiency, pursued for its own sake, is not sufficient to govern organizational activity for two reasons. First, all three models of efficiency place greater emphasis on measurable costs of things and may not place enough value on hidden or intangible costs. Business efficiency ignores hidden costs, productive efficiency can’t price them, and allocative efficiency pretends they don’t exist.
Here’s an example – Bring Your Own Device, or BYOD. This is a policy in which firms allow employees to use their own computing devices in the workplace, which also encourages their use in telework outside the workplace. There are advantages, such that businesses do not have to buy the devices and they can maintain connectivity with their workers. At first blush, this seems like a very efficient proposition. But BYOD has serious disadvantages. There have been security breaches due to theft or carelessness. Employees are increasingly found using company resources for personal use, and companies can be liable for criminal activity conducted over their systems by employees. And there are compatibility issues between personal devices and the information technology architecture — such as when employees buy more advanced devices than the architecture can handle, or when the company upgrades the architecture but employees do not upgrade their personal devices in kind. How much are these hidden costs? One study showed that a 740-person company can face costs in excess of one million dollars a year when allowing workers to use their own devices for work-related activities. The costs have steadily increased as both employers and employees demand increased capabilities.
There is a second reason why efficiency for its own sake can be harmful to organizations and employees – and that is when budget pressures drive organizations toward allocative efficiency, which is the most risk averse of the three models. The strict emphasis on matching resources to requirements discourages what is most needed when budgets are tight, greater innovation. And innovation needs flexibility, money, manpower, and the willingness to try new things and exploit success. A resource management system built to be allocatively efficient erects many barriers to such flexibility, stifling such good ideas.
Over 150 years ago, political economist Frederic Bastiat said, “There is only one difference between a bad economist and a good one: the bad economist confines himself to the visible effect; the good economist takes into account both the effect that can be seen and those effects that must be foreseen.” The same can be said for organizational leaders. Ultimately, what should be more important than just efficiency is effectiveness. What is it that the customer needs and to what extent is it satisfied? One would hope in a level economic playing field that the most effective companies would also be very efficient, but not pursuing it as an end in itself.
So, in conclusion, we should be wary of claims that centralizing business practices will make organizations more efficient. Leaders should ask, ‘Efficient, in what way? For whom? And what are the associated hidden and opportunity costs?’ If those questions have answers, then OK. Centralize. Otherwise, beware. Make sure what the organization views as efficient is aligned with the views of stakeholders, and perhaps steer the conversation back to the effective delivery of goods and services.