Lars P. Syl — Is economics — really — predictable?

There is a big difference between predicting and forecasting. Scientific theory is about causal explanation and prediction through formulating testable hypothesis that challenge the theory rigorously based on experimental evidence. Forecasting is making educated guesses based on limited and variable information information. The former applies chiefly to ergodic systems and the latter to non-ergodic, or if the system is actually ergodic, not enough it known about it to construct a rigorous causal explanation.

The ideal causal explanation is in terms of deterministic functions in which a rule applied to a single measurable input results invariably in a single measurable output. The debate over whether statistics can deliver on causal explanation is still raging, in light of the principle that correlation is not causation. For example, Einstein rejected that it could and continued to seek for a set of deterministic functions as the basis for causal explanation in physics, viewing QM as an admission of lingering ignorance about the laws of nature owing to QM being stochastic.

Libraries are full of tomes debating the details of this, but this is a rough outline to which most agree. Thus, forecasting can be "scientific" and even based on causal explanation, but it fails the test of prediction strictly speaking based on performance. The subject matter of the social sciences is more like the weather than planetary motion, and so the results are mixed. There is no ephemeris for economic cycles.

The question how sharp the line dividing prediction and forecasting may be, and this is a matter of argument since no set of criteria are universally agreed upon. Some conventional economists seek to categorize economics with the natural sciences rather than the social sciences, for example. Is that justifiable?

To understand Keynes, it is necessary to take his Treatise on Probability as a starting point.

Why is this important other than philosophically? Because humans are ideological and affected by presumptions as hidden assumptions. We tend to overestimate our level of knowledge, on one hand, and other the other, we inflate our degree of confidence.

To paraphrase Richard Feynman the purpose of science is to prevent us from fooling ourselves and we ourselves are the easiest people to fool (owing to cognitive-affective bias).

Lars P. Syll’s Blog
Is economics — really — predictable?
Lars P. Syll | Professor, Malmo University

Primer: Causality In Models — Brian Romanchuk

Since is about identifying regularity in change and developing theories that explain the causation, ideally in terms of variables and linear functions. This is a challenge even in complicated simple systems, e.g, in the natural sciences, and it is a huge challenge when dealing with complex adaptive systems in the life and social sciences. 

Complex adaptive systems are synergistic, meaning that they are greater than the sum of the parts, so that examining the parts alone is insufficient for analyzing the system as a whole, that is, the parts and their relationships.

Complex adaptive systems are also subject to emergence, that is, the appearance of properties that are unforeseeable based on analysis of the existing system. For example, systems with the capability to learn from feedback and change behavior based on learning are unpredictable based on discovery of new knowledge and its application. There is as yet no logic of discovery that formalized the process and no scientific theory that has penetrated the causality so to be be able to influence it.

Furthermore, a framework for approaching causality must be assumed and ideally defined operationally in science, as Brian does in this first sentence following:

One important consideration for indicator construction is the notion of causality (using systems engineering terminology). A non-causal model is a model where the output depends upon the future values of inputs. In the absence of access to a time machine, such a model cannot be directly implemented in the real world. In practice, a non-causal model output is “revised” as new datapoints are added to input series. The result is that we cannot use the latest values of the series to judge the quality of previous “predictions” of the model.

The use of non-causal model might be acceptable for the analysis of a historical episode, or an earlier economic regime (such as various Gold Standard periods). Since new data will not arrive, there will be no revisions....

However, causation is still an open question in the philosophy of science.•

Bond Economics
Primer: Causality In Models
Brian Romanchuk

• Causality can be defined as the "causal" connection between cause and effect, e.g., in terms of conditionality (sufficient condition, necessary condition, necessary and sufficient condition). Causality is established though a scientific theory that accounts for the connection, since "correlation is not causation."

Causation is the entire scope of the subject, which includes "causality" as just defined but is not limited to it. There are ontological and epistemology issues regarding causation that are not settled. 

Generally speaking, modern science assumes 1) ontological monism in assuming naturalism, that is, that "everything" can be explained by natural causes as observables (as in a theory of everything). It also assumes 2) epistemological realism in the sense that the mind (subjectivity) is the mirror reality (objectivity) when the scientific method is correctly applied. 

However, these assumptions regarding the framework for gaining knowledge are more presuppositions than stated assumptions. Philosophy of science attempts to bring clarity to this by examining the various issues that arise.

Brian Romanchuk — Representative Agent Macro And Recessions

J.W. Mason kicked off the latest skirmish in the never-ending macro wars with his Jacobin article "A Demystifying Decade for Economics." (Note: at the time of writing, the article was taken down until its publication in Jacobin.) This prompted a Twitter debate about representative agent macro, which eventually led to this Beatrice Cherrier article on heterogeneous agent models. In my view, the debate about representative agent models is a red herring. Mainstream macroeconomists main skill is in framing debates in a fashion that is congenial to the mainstream; however, the preferred framing leads to dead ends. My current research focus is on recessions, and although I have not gone too far in refreshing my survey of mainstream macro, the value of mainstream macro theory in this debate is limited....

Bond Economics
Representative Agent Macro And Recessions
Brian Romanchuk

Brian Romanchuk — The Kalecki Profit Equation And Forecasting

Having run through the Kalecki profit Equation (link to the first part of a two-part primer), I just want to make some brief remarks about how it ties into the notion of forecastability (description). Should we be able to expect to forecast the business cycle?

Bond Economics
The Kalecki Profit Equation And Forecasting
Brian Romanchuk

Brian Romanchuk — Forecastability And Economic Modelling

When most people think about macroeconomics, what they want is the ability to forecast economic outcomes. However, economists' (of all stripes) reputation as forecasters is not particularly high. My view is that this is not too surprising: what we want forecasters to accomplish is probably impossible. (I am hardly the first person to note this, as variants of this idea go back at least to Keynes; I could not hope to offer a history of this idea.) However, I think if we want to approach macro theory formally, we need to formalise the notion that outcomes cannot be forecast, which means we need to define non-forecastability formally.

This article gives one potential definition of forecastability, and then applies the concept to a simple stock-flow consistent (SFC) model. It should be noted that these are my preliminary thoughts, and I believe that the definition will need to be refined.…

This is one of the key questions in philosophy of economics, as well as philosophy of science and philosophy of social science.

The purpose of science is to provide as general an explanation of data ("data" means "the given.") The data set is determined by the nature, scope and scale of the subject matter being explained. This is accomplished through modeling, both conceptual and mathematical.

Prediction comes in with respect to testing outcomes of theoretical models, using hypothesis the theory generates and carefully designing testing apparatus.

The difference between philosophical speculation based on reasoning and science is that scientific reasoning can be tested by subjecting hypotheses to disconfirmation, since a general statement (theory) is contradicted by a single false instance.

No amount of true instances can definitely confirm a general statement that is not a tautology and therefore empty of content about how things stand in the world. This is the case logically, even though we call theoretical assumptions that are well confirmed by hypothesis testing "scientific laws."

Science is always tentative on the next outcome unless it is established that all the factors involved are known to be true based on observation. A logical argument is sound if and only if the logical from is valid and the premises are true. Then the conclusion necessarily follows.

This implies that a great deal of that which is considered scientific is speculative. That is to say, it is not science but philosophy.

There is nothing wrong with philosophy. Not everything is explained by scientific reasoning. The questions involving key fundamentals of life and reality have not been answered using scientific reasoning. They are "the enduring questions" that are the domain of philosophy. When methods are developed to answer such questions using scientific reasoning, then they become the subject matter of science.

I don't wish to give the impression that fundamental issues in philosophical method are resolved. They are not for the simple logical reason of lack of criteria that are universally agreed up. But the above more or less summarizes what is generally accepted practice based on logic., even though there are issues in the foundations of logic, too.

Hopefully, Brian's post will contribute to getting economists thinking more  about the foundations of their field and doing this more carefully.

Bond Economics
Forecastability And Economic Modelling
Brian Romanchuk

See also

My point is merely that forecasting is not the same as modelling, nor the same as telling a good story.

Stumbling and Mumbling
On thin predictions
Chris Dillow | Investors Chronicle

See also

There are many arguments for the use of models in economics, including notions of rigor and transparency, or that models can help you to see relationships you otherwise might not have expected. I don’t wish to gainsay those, but I thought of another argument yesterday. Models are a way of indexing your thoughts. A model can tell you which are the core features of your argument and force you to give them names. You then can use those names to find what others have written about your topic and your mechanisms. In essence, you are expanding the division of labor in science more effectively by using models.

Austrian economists of typically better at examining the foundations of economics, probably since Mises and Hayek were philosophers. Austrian economists have also written at great about the foundations of liberalism as a social and political theory.

Marginal Revolution
Models as indexing, and the value of GoogleTyler Cowen | Holbert C. Harris Chair of Economics at George Mason University and serves as chairman and general director of the Mercatus Center

Adam Shaw — Why economic forecasting has always been a flawed science

So how do you make good predictions? I met “superforecaster” Michael Story, who was ranked 18th best among the 20,000 people who formed the Good Judgment team. The team took part in a competition conducted by the US intelligence community to find the world’s best forecasters. Launched in 2011, the four-year contest required the group to provide forecasts on 500 questions ranging from the future for oil prices to the financial outlook. The Good Judgment team won the tournament, reportedly outperforming even professional intelligence analysts with access to classified data.

The grading of their volunteers’ forecasting abilities was key to why Good Judgment did so well. People knew how well they were performing and were driven to improve. They were also encouraged to correct their biases and alter their world view amid changing circumstances. This self-analysis and willingness to adapt, they believe, was crucial to the team’s success.…

Conclusion.

The Good Judgment team believes part of the problem is that we misunderstand the science of forecasting and look to the wrong people for predictions. If we want to know what’s happening to the economy, we think the obvious thing to do is ask an economist. But Storey says that may be the wrong approach. Forecasting is an art that is separate from the need to have specific subject knowledge. The people who were best at predicting the Arab spring, he said, were not Middle East experts. They were people who studied eastern Europe and had seen similar patterns develop there. We don’t need subject experts, we need people who are great at forecasting anything.

The Guardian
Why economic forecasting has always been a flawed science

Adam Shaw

Daniel Little — Predicting, forecasting, and superforecasting

Issues in forecasting social events versus predicting natural events.

Understanding Society
Predicting, forecasting, and superforecasting
Daniel Little | Chancellor of the University of Michigan-Dearborn, Professor of Philosophy at UM-Dearborn and Professor of Sociology at UM-Ann Arbor

Nate Silver — Models, Models, Everywhere

We assign a lot of bandwidth to the FiveThirtyEight presidential forecast model, but that doesn’t mean it’s the only thing you should look at to get a sense for where the election is headed. So it’s time for a quick gut-check. How does the FiveThirtyEight model — where Barack Obama’s probability of winning the Electoral College recently ticked above 70 percent — compare against alternative means of forecasting the election?

There are a number of other statistical forecasting systems, most of which rely on polls, economic variables or some combination thereof.

I tracked down about every one of these models that I could find, subject to the condition that it couched its forecast in probabilistic terms (or that it was well-documented enough to allow this to be inferred with relative ease, like from the standard error that the model stated).

In my view, it’s in estimating the uncertainty in a forecast where most of the challenge and intrigue lies. To paraphrase Charles Barkley, any knucklehead can make a point prediction — but it takes brains to calculate a confidence interval.

Read it at The New York Times | Five Thirty Eight
Models, Models, Everywhere
Nate Silver