Week 2 - Causality and Applications to economics

Question 1

What is Aristotle’s doctrine?

Answer 1

Science is “scire per causas”, i.e. to know through causes. Science is “etiology”.

We have science only when we can reach the causes, or first principles of things. This knowledge has to be necessary, not hypothetic

Aristotle is interested in the structure of the concept of motion and the condition of its possibility

Question 2

What are Aristotelian doctrine of causes

Answer 2

1. Material cause: internal constituent, like the material of a certain object
2. Formal cause: form or model, i.e. definition of a certain being
3. Efficient cause: the element producing change
4. Final cause: the scope of the action

Question 3

What was Galileo’s view on modern science?

Answer 3

According to some authors modern science is characterised by the rejection of the search for causes, by substituting to the latter the search for connections.

According to Galileo instead, there is a different kind of relationship between reason and experience which looks for the prediction of facts.

Question 4

Explain the Deductive - Nomological Method

Answer 4

A scientific explanation has to show how a singular fact is an instantiation of a fundamental law

What has to be explained (explanandum) is deduced from some true statements including at least one law (explanans)

Question 5

What is the problem with the deductive-nomological approach?

Answer 5

Problem: not all logical deductive inferences can be considered as laws. Example:
Nobody who takes birth control pills gets pregnant
George takes birth control pills
George doesn’t get pregnant

Question 6

Explain the D-N model?

Answer 6

Deductive inference:

1. Law
2. Initial condition
3. Explanandum

Without laws there are no explanations and no scientific prediction ie no science

Explanation of the past and prediction of the future constitutes science (symmetry thesis)

Any good explanation must indicate the causes!

Question 7

Explain the empiricist view?

Answer 7

Inductive Inference:

1. Science collects data concerning variables
2. Use these data to infer general laws

Question 8

Explain Mill’s practical problem?

Answer 8

The world is too complicated to derive predictions, with a simpler world (in laboratories for example) it could be easier to obtain general laws.

The practical problem of induction is related to our limited knowledge

Economic phenomena differ among them and it is not possible to collect events that are completely similar.

It is difficult, if not impossible, in social sciences to isolate phenomema

There are too many correlations to allow to select true causes.

Question 9

Explain Mills approach?

Answer 9

According to Mill, causality, i.e. “anything that has a beginning has a cause”, is at the basis of any human knowledge

Mill says that a simple regular sequence does not identify a causal relationship, and he distinguishes between necessary conditions (conditions without which there would be no consequent but not sufficient by themselves) and sufficient conditions (those which can bring about an effect by themselves)

Question 10

What was Hume’s challange?

Answer 10

Hume wants to replace causal laws with regular sequence of events, thus denying the necessity of causal relationships

When we say that a causes b we are assuming:

1. There is a class of events A including a and B including b
2. Events in A are regularly followed by events in B

Question 11

Explain Hume’s theoretical problem?

Answer 11

Let’s suppose we have found a strong correlation so far in any experiment or in any experience we had, how do we know it will continue to hold in the future?

According to Hume we just believe in the persistence of laws, i.e. we are assuming uniformity and regularity in nature
Vicious circle: to justify laws of nature we use the principle of uniformity, but the latter is justified on our past experience!

Question 12

Explain Hume’s Challenge?

Answer 12

According to some philosophers the very idea of regular sequence presupposes the existence of a law distinguishing regular and accidental sequences.

Notice that Hume doesn’t deny that things are brought about by causes, he only denies we can prove or have intuition of this.

Causal relationship is not a logical (necessary) relationship
That Nature follows a certain uniformity is not a logical truth but is only sustained by habit (custom)

Laws are just a tendency or habit of our mind to accept some correlations and regularities as laws

According to this view all the scientific experience in any field is based on an irrational belief.

The problem is not that it is difficult to find causes, but that we cannot associate the same effects to the same causes.

Question 13

What are the consequences of Hume’s problem?

Answer 13

Instrumentalism: Laws are merely useful instruments whose value relies not on its truth but on its capacity to predict.

Confirmationism: Laws only express probabilistic knowledge (i.e. only express a certain probability for an effect to occur given a cause)

Question 14

Explain Kant’s reply?

Answer 14

According to Rationalists causality can be grounded on principles independent from experience, whereas Hume grounded it on psychological constitution of mind

Kant considers Hume’s challenge as a sound awakening against any form of dogmatism, and concludes that a role for causal laws can only be founded in the mind that imposes its laws to reality

Kant asserts that human mind is endowed with a transcendental structure which imposes its laws upon the objects

According to Kant what we can know is not reality in itself but our representation of it.

Our knowledge starts with the intuition of the object which is possible through the a priori forms of time and space

Causality refers to the necessary connection of a cause with an effect, however Kant agrees that its justification is not logical but is a priori, i.e. a law of mind. It is pure concept or category which is applied to our representations.

Question 15

Explain the difference between deduction VS casual explanations

Answer 15

Moreover causal relations (like explanations) are asymmetric whereas deduction is symmetric

Deduction. The height of a flagpole intercepting the sun causes the shadow, thus if you know the height of flagpole, the position of the sun and the appropriate physical law you can deduce the length of the shadow. Viceversa if you know, the length of the shadow, position of the sun and the law, one can deduce the height of the flagpole.

Causal Explanation. But whereas the position of the sun causes the shadow, the shadow does not cause the position of the sun, thus knowing the length of the shadow doesn’t help explaining the position of the sun.

Question 16

Explain why cofounder variables are a threat to economics?

Answer 16

Cofounder variables pose a serious problem in economics because in many cases whether the correlation between X and Y can be measured, the cofounder cannot. In medicine, for example it wasn’t clear (until recently) whether smoke caused lung cancer or they were both expression of a common predisposition.

In many time series, correlation arises because the series are both non stationary and follow the same path but there is no connection among them.

Question 17

Hume… Again.

According to Hume, X causes Y if:

Answer 17

1. X is universally associated with Y
2. Y follows X in time
3. X and Y are spatio-temporally contiguous (no gaps)

In other words causal relationship is equivalent to constant conjunction. E.g. Irresponsible bank behaviour as a cause of the financial crisis.

However these conditions that define causality as constant conjunction are neither necessary nor sufficient to explain an event.

Question 18

What are the flaws in Hume’s account of casuality

Answer 18

Causes may not be universally associated with their effects. E.g. Irresponsible bank behaviour may not lead to a crisis because of contrasting events

Not all effects follow their causes. In many cases, the effect is simultaneous to the cause. E.g. a ball put on a cushion creates simultaneously a hollow.

Causes may act after a lag or at a spatial distance. E.g. an industrial policy against monopoly power today may reduce incentives to innovations with effects after many years.

Hume’s 1-3 are not sufficient to be a cause, if all students in a cohort develop symptoms of stomach disease one after another, this doesn’t mean that there is a causal relationship, because it may well be that they have all eaten the same gone-off food.

Question 19

What are tendency laws?

Answer 19

According to Mill, economic laws do not express universal regularities but rather causal tendencies.

If laws are taken to mean universal statements, then many laws would be false. Even a physical law such as the law of falling bodies can easily be disproved if we compare a feather and a marble (this is due to air resistance). This is even truer for economic laws.

However if we isolate the falling object by eliminating air resistance the law holds true.

This means that laws hold under special conditions (i.e. ceteris paribus)

Ceteris paribus means “all other things being equal”, but in some cases it is used to say “all other things being absent”
Now, a tendency law is a law that holds true when disturbing factors are absent (the causal relationship is isolated from disturbing factors).

In this sense laws are universally true if considered as tendencies (i.e. when the disturbing factors are taken out).

A second characteristic of tendency laws is that they exert a power, i.e. there are factors that make things happen
Now the problem is to understand what happens when we cannot isolate phenomena, i.e. when “other things are not equal”

According to Mill, in economics different forces compound together to establish the final direction of the effect.

A third characteristic is that they produce an effect even when disturbing factors are present.

Let’s take the quantity theory of money: MV=PT, or in its simplest form M=kP, i.e. M is proportional to P.
This may be interpreted as a causal tendency because:
The change in money causes a change in P.
This is only true if disturbing factors such as productivity shocks and raw material shocks (etc.) are taken out.
Money has an effect on prices even when disturbing factors are present and even when the overall result may be the opposite of what is predicted by the Quantity theory (P may decrease because productivity more than compensate the increase in M).

Question 20

A third characteristic of tendency laws is that they produce an effect even when disturbing factors are present. Provide an example of this…

Answer 20

Let’s take the quantity theory of money: MV=PT, or in its simplest form M=kP, i.e. M is proportional to P.
This may be interpreted as a causal tendency because:
The change in money causes a change in P.
This is only true if disturbing factors such as productivity shocks and raw material shocks (etc.) are taken out.
Money has an effect on prices even when disturbing factors are present and even when the overall result may be the opposite of what is predicted by the Quantity theory (P may decrease because productivity more than compensate the increase in M).

Question 21

What are the problems with tendency laws?

Answer 21

In physics, if one compounds the forces of the tendency law and the disturbing factor, one gets the overall effect. In economics, the laws of the disturbing factors may not be known (e.g. you may not know the effect of productivity on P)
When there are many disturbing factors this problem is even more serious
A counterfactual interpretation maintains that tendency laws explain an outcome when all the disturbing factors are at work but the main factor is not. E.g. The QT doesn’t explain the actual P, but why P is not what it was predicted to be (i.e. because of disturbing factors).