On the natural side, are not the observables, theoretically if not actually, the interesting questions in natural sciences?
Look at your list.
"Human nature". What is relevant to providing cause and effect explanations are regularities of human behavior. As an evolved social primate, there will be innate biases, and to the extent that they are relevant to regular behavior, they will be subject to scientific observation. And of course, those behaviors include the internalization of the regular rules of behavior in the social contexts that we encounter, which leads to institutions that are themselves observables in their rules of behavior and the folkviews regarding those rules of behavior held by their participants.
Any unobservable element of human nature over and above that is beside the point in providing cause and effect explanations human behavior.
"Motivations". Mainstream economics has a strong reliance on unobserved and unobservable utility preference maps as the motivation for the incessant decision making followed by performance that is its unit of analysis. But that is required for the scientific study of human behavior, but only for application of that particular unit of analysis of evaluation followed by decision followed by performance. Replace the unit of analysis with social transaction followed by performance, and then to the extent that motives are relevant to the transaction, they are observables.
"Inherent attributes". Inherent attributes are either observables or excuses in lieu of study of regularities of human behavior.
"Justice". The various rules of appeal to various internalized models of justice are quite observable, as is the regularity that models of justice are formed and internalized. Often those models of justice have a folkview that the model of justice itself is intrinsically valid rather than socially grounded. There is an "unobservable" there if we attempt to find that intrinsic grounding of some particular model of justice, but that is the same as the unobservability of the orbits of the sun and the planets around the earth ... we cannot observe what is not there.
"Equity". This is, of course, a specific facet of some specific models of justice.
"Well being". Basic needs of humans can be identified, without difficulty. The idea of some generic unstructured quantity of "well being" is, of course, more of the long falsified utilitarianism of mainstream economics. Those aspects of well being that are observable and identifiable are precisely those aspects that social science can address. I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.
Most economist wouldn't agree that your definition describes the field, currently or ever. Certainly the classical economists, Smith, Ricardo, and Marx wouldn't have. Economics was a branch of moral philosophy, and material provisions were always proxies for broader concepts of justice and ethics. That hasn't changed. Economics is the study of how to organize society in ways that provide for more justice and greater well-being. It is a normative discipline, in that economists are expected to prefer social organizations that provide greater justice and greater well-being to those that don't.
Regarding your criticisms of my partial list of unobservable phenomenon that make up the basis of what are concerned "interesting" research questions in the social sciences, I refer to you the seminal work on the topic of causality by the eminent computer scientist and mathematician (not an economist) Judea Pearl (http://bayes.cs.ucla.edu/BOOK-2K/) whose work on the science and philosophy of causality is considered foundational in both the natural and social sciences.
(As a sad aside, Dr. Pearl is also, quite tragically, better known outside of academia as the father of David Pearl, the Wall Street Journal reporter who was beheaded on international TV by his captors in Pakistan in 2002.)
I'll take just one topic from your list, because it is the biggest problem in all of social science research, from psychology to sociology to economics -- inherent unobservable attributes. A really good example that comes from the the controversial arena of teen pregnancy, sex education, abortion, and life outcomes. Up until recently, most research showed, unsurprisingly, that girls who became pregnant at a young age had worse life outcomes than girls who didn't when categorized into those two groups and other factors were sufficiently controlled for in statistical models of the relevent relationships. However, that still never satisfied the on-going problem in statistics of missing variables -- what didn't get included, and how that might have biased the inference of causality between pregnancy and poor outcomes. What is a missing variable? It's almost always some unobservable phenomenon for which it is impossible, or nearly so, to find data. In the case of pregnant teens, the question is, "Is there something else about these girls that both causes them to get pregnant early and also causes them to have worse life outcomes?" That is, is there some unobserved characteristic in some girls that causes them to BOTH get pregnant and have other problems unrelated to the pregnancy? If so, pregnancy can't be the cause of poor life outcomes, and policies and social mores focusing on preventing teen pregnancy are probably misplaced.
Well, a few labor economists (http://jhr.uwpress.org/cgi/content/abstract/XL/3/683)came up with a novel solution to the problem by looking at the US longitudinal survey data and grouping teens into different categories -- those who got pregnant and gave birth, and those who got pregnant but didn't give birth. Their finding was surprising to them. Girls who gave birth had as good or better life outcomes after 10 years than girls who didn't give birth. This provides some evidence about causality -- that pregnancy cannot be said to cause poor life outcomes in and of itself. Rather there is some still unidentified characteristic(s) inherent in some girls that explains both getting pregnant early and having poor life outcomes. Girls should therefore not expect poor life outcomes if they have a child as a teen, ceteris paribus, but they should wonder if there is something else about them that might still lead to poor outcomes whether or not they carry a child to term.
There is always a tradeoff between modelling a process correctly (ie without missing variables, entirely in terms of what has actually been observed and nothing else) and adapting an off the shelf model while hoping it will work out.
The reality is that the substantial effort required to model a process from scratch is not justified in nearly all cases. The few standard (statistical and physical) models that were developed from scratch since the Renaissance have been reused and extended many times, to the point that the cost of using them is merely a few years of university education. -- $E(X_t|F_s) = X_s,\quad t > s$
You're quite right that in most cases the substantial effort to make a new model from scratch is unjustified, and this gets to the heart of Bruce's issue with the institutionalist critique. We're biased by the nature and necessity of our circumstances to accept and build on the models and mistakes of others. Which means that we're likely to miss important things that don't fit the models, such as unobservable phenomenon.
However, even in making a model from scratch there is still the non-trivial issue of unobservables. This is the real problem that most economists, as well as many other social scientists, and even medical researchers, struggle to answer: "What WOULD have happened if X were true instead of Y?" -- a counterfactual, in other words. That's how causality is best inferred and how statistics is used to find the answer, but doing so is really hard work because counterfactuals are, by definition, unobservable, which means that a better theory makes all the difference.
A missing variable is a problem IF there is, in fact, a variable missing from a statistical model. The problem is that often only theory can tell you if it is missing or not -- not anything in your model itself. This means that you'll never know if it's missing if you haven't thought sufficiently about your problem.
That is why you make control experiments.
This is the real problem that most economists, as well as many other social scientists, and even medical researchers, struggle to answer: "What WOULD have happened if X were true instead of Y?" -- a counterfactual, in other words. That's how causality is best inferred and how statistics is used to find the answer, but doing so is really hard work because counterfactuals are, by definition, unobservable, which means that a better theory makes all the difference.
And this is why you do double-blind placebo-controlled clinical trials.
- Jake If you only spend 20 minutes of the rest of your life on economics, go spend them here.
Seen from the perspective of an anthropologist and a historian, the history of the field of economics has been one of a long flight from the complexity of reality into a warm numerical cocoon of fantasy.
We don't understand why people do things? Let's just pretend that all people try to maximize their utility. We can't model asymmetrical information in an adversarial market situation? Let's pretend everybody knows everything.
The fact that these models and this way of thinking encouraged people to look away from the structural problems causing poverty, to look away from gluttonous parasitism at the top, to look away from the role of hierarchy and power in defining the market situation, was merely a bonus. It's so much easier to just blame poor people for failing to maximize their utility.
