What is value?

Some foundations and a new definition

About 20 years ago, a scrap metal dealer in the midwestern US bought a jeweled egg for $14,000 at a flea market with the intention of melting it down and selling it for the gold from which it appeared to be made. But instead of melting it right away, he put it on his mantelpiece at home. It was nice to look at. So he looked at for a few years.

A few times, he offered to sell it to other metal dealers, who told him that the gold content was worth less than what he had paid for it. He thought it was more valuable than that, but he didn’t really have any evidence. Besides, he liked it.

In 2012, he was at last seized with curiosity and Googled the word “egg.” In the results, he learned for the first time what a Fabergé Egg was, and even found an article in the Telegraph that pointedly asked him “Is this £20 million nest-egg on your mantelpiece?”

The scrap dealer contacted an actual expert, who flew in from London, inspected the Third Imperial Fabergé Egg, and quickly made a deal probably in the tens of millions of USD. The details of this story are fuzzy because the scrap dealer has managed to maintain his anonymity to this day and we don’t even know who currently owns it.¹

As etymology would suggest, evaluators find the value of things. But what is value? Before I give my answer, I’d encourage you to pause and try answer this question yourself.

As you might be able to predict, there are several plausible answers to this question. Let’s start with what we typically say that evaluators do and then move on to some angles on value that we don’t typically discuss.

The job of evaluators is to determine the merit and worth of things (evaluands).

Michael Scriven defines “merit” as: “The ‘intrinsic’ value of evaluands, as opposed to extrinsic or system-related value/ worth. .”²

His definition of “worth” begins: “Usually refers to the value to an institution or collective, by contrast with intrinsic value, value by professional standards, or value to an individual consumer (merit).”³

His example of how these terms relate to one another is that “the merit of researchers lies in their skill and originality, whereas their worth to the institution that employs them) might include the income they generate through grants, fame, bequests, attracting other good faculty and students.”⁴

The common characteristic of both merit and worth is that they concern value. The primary difference between them is that merit is intrinsic value and worth is instrumental value. By defining evaluation in terms of both merit and worth, Scriven and other evaluators take a philosophical position on intrinsic and instrumental value.

Intrinsic and Instrumental Value

Many philosophers claim that something has intrinsic value if it is good “in itself,” independent of any external factors. For instance, in some ethical theories, happiness, beauty, or the good will (as in Kant’s moral philosophy) is intrinsically valuable. The idea is that these qualities are worthwhile not because they lead to something else, but just because of what they are. This is the value equivalent of “we assert these truths to be self-evident…” They simply are valuable. I think pleasure falls into this category – it doesn’t make a lot of sense to ask why it is valuable for a meal to taste good. A tasty meal just is more valuable than a bland one.⁵

By contrast, something has instrumental value if it is valuable only as a means to an end. Money, for example, is typically held to have instrumental value because its worth derives from what it can buy, not from any inherent “goodness.” Aristotle was fond of asking people why they wanted valuable things. He noticed that he would go through a few layers of instrumental value before hitting a bedrock intrinsic value: you want a career… because you want money… because you want freedom… because you want to lead a fulfilled life.

By insisting that evaluation is about determining both merit and worth, Scriven and other evaluators are taking the position that both intrinsic and instrumental value are legitimate forms of value. The evaluand can be good in itself and good for other purposes. This means that, if we are evaluating a medical clinic:

1. It can have merit if it is well-run, provides a high standard of care, and is a good place to work.

2. It can have worth if, each year on average, it saves 100 lives, adds 500 quality-adjusted life years to the community, and generates a net value of $10 million to the local economy.

The classic definition of evaluation says that these are both legitimate kinds of value.

Realist and Anti-realist Value

Let’s return to our standard account of what evaluation is: determining merit and worth. By saying that people can “determine” merit and worth, evaluators are taking another philosophical position. Evaluators treat value as something that can actually be found. For example, it makes sense to ask “What is the present value of a Rimac Nevera (the Croation electric hypercar)?”⁶ because we believe that there is an answer to this question. If we didn’t believe that there were possible answers to this question, we probably wouldn’t say “determine” – we would use a verb like “stipulate” or “assert” or “socially construct” to describe what evaluators are doing.⁷

This brings us to another philosophical distinction, that between a realist and and an anti-realist conception of value. Value realism posits that values exist autonomously of our thoughts or feelings about them. According to this view, when we say “this clinic is good,” we are stating a fact about the world. Other classic theories, like emotivism or expressivism, argue that value judgments are expressions of our feelings or attitudes. From this perspective, saying “this clinic is good” is less about stating a fact and more about expressing approval: “hooray for the clinic!” or “I commit to acting as though this were a good clinic.” One version of anti-realism – error theory – says that our value discourse is systematically mistaken because values simply don’t exist.

In our heavily quantified world, it isn’t hard to imagine being realist about at least some instrumental values. You negotiate for your wages, you shop online, maybe you hold stocks. It’s obvious to a person living in the 21st century that there are external forces that shape the value of things such that you can go find out how much certain things are worth. However, it’s a bigger philosophical challenge to be a realist about intrinsic value. What procedure would we use to find out the value of a pleasure, for example?

There is one common answer for this question, but it doesn’t satisfy everyone. Most economists would say that that people have preferences, even between things on which they would have a hard time putting a price. Today I had a choice between going to the movies and writing – I chose writing. I have no idea what exact value I would place on the experience of going to the movies (particularly because every movie is different) and the experience of writing (also because everything I write is different). However, my revealed preference today was for writing. Most economists would say that this means I value writing more right now, and we can at least put writing and movies on an ordinal value scale for me. Even if people can’t tell you what they value, they will show you.

Value and Utility

Some sharp-eyed readers may have noticed that the procedure I just explained for understanding revealed preferences is the classical method of determining utility, not value. Indeed, I believe the two concepts are closely related. To some people, this will sound strange, but this is probably because they are thinking of utility as a synonym for “useful.” In decision theory, utility actually doesn’t refer to how useful things are – rather, it axiomatically refers to our preferences. In classic decision theory it can be shown that, if I can order your preference into indifference classes (strata corresponding to options between which you either either prefer A to B or are indifferent) then I can also use probability to determine the intermediate points of utility between those options we have using a lottery. The exact proof of this goes beyond the scope of this post, but the upshot is that we can use ordinal preference rankings to derive interval utilities – all we need to know is how much you prefer each option when confronted with a series of choices.

However, there are some essential differences between utility and value that we need to appreciate if we are to understand the real challenges of determining the latter. Utility is defined entirely in terms of preferences, while value is usually thought to exceed mere preferences. If value were reducible to utility, it would be incoherent to make a statement like “she wanted sapphires more than diamonds, even though diamonds were more valuable.” Utility is allowed to vary idiosyncratically from one person to another, but value is generally supposed to have some social dimension. Even if someone personally prefers Dan Brown to Marcel Proust, it makes sense for me to assert that the works of Proust have greater value. This is not merely an appeal to popular opinion (though I’m afraid Proust would lose in that event) but rather an appeal to literary criteria like originality, thematic depth, facility with language, and so forth. If you walk into a store and buy The DaVinci Code instead of the first volume of La Recherche, assuming a negligible difference in price, you have revealed a preference and thus a utility, but this not the same things as the value of these books. Perhaps if needed to decide which one to put in the local library or a time capsule, you would choose differently. The reason utility can never stand in directly for value is that utility has the characteristics of an operationalist definition – it is an abstract concept defined stipulatively by the procedures used to measure it, rather than its characteristics.

Supply and Demand

One aspect of conceiving of value as a cousin of utility is that it makes value a function of demand. The more we prefer something, the more utility it has for us. It’s possible to conceptualize value this way. Why are diamonds so valuable? Because so many people want them, the classic response goes. The definition of “value” might be the extent to which something is wanted or what we are willing to give up to get something. Indeed, this is Georg Simmel’s classic definition in The Philosophy of Money (1900).

There is another famous way to look at the question of value which we should consider that flips this conception on its head: the labor theory of value. This is a supply-side version of value, in which value is invested in things by their creators, and, as Aquinas put it “value can, does and should increase in relation to the amount of labor which has been expended in the improvement of commodities.” Until the late 19th century, the labor-theory of value (LTV) was the default economic theory, expounded by Adam Smith, David Ricardo, Benjamin Franklin, Pierre Joseph Proudhon, Lysander Spooner, and virtually anyone who thought about economics. Today, Marx generally gets sole credit for the theory. In the LTV, value is something that people put into things, and things are more valuable when more work goes into them.

The most obvious objection to the LTV, that the theory rewards unproductive labor, is handily dealt with by Marx. He explains that really we are paying attention to how much effort the average worker takes to make things, rather than the individual worker. However, I think the LTV is wrong for other reasons. As a theory of prices the LTV obviously doesn’t work very well, since prices are subject to high variance for goods that are made under essentially identical conditions (think of the brand name versus generic distinction). As a theory of value, it puts us in the untenable position of claiming that any two things are equally valuable if they required the same amount of labor to produce. If I spend 50 hours making a report that turns out to be useless and 50 hours doing a report with data analysis that saves a nonprofit a million dollars, these two products are equally valuable according to the LTV. In order for us to take the LTV seriously as a theory of value, it needs to say something about the quality of products.

As I mentioned above, things changed in economics in the late 19th century. Thinkers like Léon Walras and Alfred Marshall developed the now-familiar supply and demand model, which explained prices as outcomes of interactions between labor inputs and consumer behavior. Even if two goods are produced with identical labor inputs, differences in consumer tastes, scarcity, and competitive pressures lead to variation in prices. If two goods require the same amount of labor but differ in quality, the equilibrium prices will differ because consumer demand responds to perceived quality, usefulness, or scarcity. The market adjusts so that the high-quality report’s price is higher and the low-quality report settles at a lower price.

Now, this is a theory of price, not value, but I think it does have some implications for our discussion once we remember that price is an indicator of value. That is, prices are noisy streams of data that carry signals about value. The lesson for understanding value is that value is related to both inputs and outputs. However, I believe that this relationship might not be the usual “efficiency” relationship of the ratio of outputs to inputs. Rather, I think that the value relation might be better described as a form of throughput. More interestingly, I don’t think the critical inputs to the value equation are actually labor.

Throughput

These days, I generally see the world in a symmetric relationship of humans and non-humans bound together in networks of materials and meaning. Action in a such a world involves bringing things and people into the network and transforming them so that they can enact other transformations on the network. I posit that this is how social change happens, and there is nothing above or outside this.

What if value is the ability to materially and semiotically transform other things and people in a network? This would be a broader umbrella conception that covers the LTV, since what workers are doing is enacting a material transformations on some commodity. Explicitly thinking in terms of semiotics (meaning) allows us to account for aesthetic modes of production as well. Adding in the requirement that the evaluand be able to transform people and things would move us beyond the supply and demand dichotomy from economics. The ability to transform nodes in the network is what distinguishes the useful report from the useless one in the example above.

Labor is certainly part of the network that is being transformed by the evaluand, but this is rarely the part that most directly contributes to value. Rather, we mostly value things because of how they transform parts of the network not directly related to the labor used to create them. When you buy medicine to treat an illness, you are trying to transform biochemical processes in your body into other processes – the value of the medicine come from the ability of medicine to enact this transformation, not the cost of the medicine or the labor involved in making it. (Think of all the nigh useless tonics that humanity has labored to produce throughout history and how much they cost.)

This is how I would start to describe value in terms of actor-network theory, but so far I’ve still neglected the most important limiting factor in economics – time. Things are more valuable if they can enact material-semiotic change with higher velocity. A medicine that could reduce the duration of symptoms from the common cold by 50% (all other things being equal in terms of side effects) would be more valuable than a medicine that reduced them by only 10%. You prefer to read a newsletter that gets to the point instead of rambling around interminably. Great artistic works are those that are relentlessly attracting and producing meaning. These evaluands have a high throughput of material and semiotic resources; they transform the material and mental world at high velocity.

The Three Big Questions

At this point it is helpful to distinguish between three questions, which are often conflated:

1. What is value?

2. Where does value come from?

3. How do we learn about value?

The first question is the topic of this essay. The second question is about causality. The third question is about epistemology. Conflations between questions #1 and #2 seem to lead people towards “supply-side” LTV, since these the LTV clearly says where value comes from – the labor of workers. Conflations between #1 and #3 seem to lead people towards “demand-side” theories of value, since market prices and revealed preferences are good ways to learn about value. However, these theories suffer from the operationalist fallacy of defining an abstraction in solely terms of the procedure used for measuring it. Money is a way of measuring value, not value itself.

My current working definition of value as an evaluator is that value is a multidimensional attribute of objects and processes, with key dimensions of value including the throughput of the object’s material and semiotic resources and its equilibrium price.⁸ Value is ontologically subjective and epistemically objective, meaning that it is mind-dependent but knowable. A consequence of this definition is that evaluation is fundamentally an exercise in measuring value, with all the attendant philosophical challenges of actual measurement.

Conclusions

Some people think that philosophy never gets anywhere, so when I talk about philosophy, I like to be very clear about what we’ve learned from our inquiry:

• The standard account of value in evaluation is pluralist. It embraces both intrinsic and instrumental forms of value.

• The standard account of value in evaluation is realist. It presupposes that value is at least partly autonomous of our attempts to measure it or make assertions about it.

• Value is not synonymous with utility, because utility is individual and entirely a function of preferences, while value is social and not defined axiomatically by our behavior towards valuable objects.

• I think value is a multidimensional attribute of objects and processes, with key dimensions of value including the throughput of the object’s material and semiotic resources and its equilibrium price.

• Value is ontologically subjective and epistemically objective, meaning that it is mind-dependent but knowable. This means that it is potentially measurable but that doing so is challenging. This also means that that the fundamental activity of evaluation is measuring value.

1. The Fabergé egg shown in the picture is the Gatchina Palace egg, which is not the one that was found by the scrap dealer – it’s just my favorite one.

2. Evaluation Thesaurus, p. 227

3. Evaluation Thesaurus, p. 382

4. Evaluation Thesaurus, p. 227

5. One particularly strong version of the view that intrinsic values are the most important kind of value is naturalism. This is the idea that value can be reduced to natural properties like pleasure, satisfaction, or biological functioning. As a discipline, evaluation does not seem to pay any heed to naturalism. I think this is because we don’t get any mileage from this kind of reductionism. For example, we might give an organization positive marks for being efficient without needing to trace this efficiency back to some natural good like pleasure. Most evaluators accept that there are a wide variety of ways that things can be valuable, and tracing particular instrumental values back to a small set of natural goods actually makes our work less specific.

6. North of $2 million, I hear.

7. If you do this, please give up being an evaluator and send me your client list.

8. Why include equilibrium price as a dimension of value? Equilibrium prices have special characteristics that give use strong social signals about value. Absent market distortions (which are extremely common), equilibrium is reached when the price adjusts to a level where the amount demanded by consumers exactly equals the amount supplied by producers. At this “market-clearing” price, there is neither a surplus nor a shortage. Once we arrive at equilibrium, prices tend to be stable because any deviation from this balance creates forces that push the market back toward equilibrium. The rate of profit falls, which means that the price gets very close to the cost of production. By paying attention to equilibrium prices and market conditions, such as distortions, we can make a lot of defensible inferences about value.