Lecture 11 – answer to possible exam question
Describe one case of your choice from economics, finance or business, where somebody is making a
causal inference. Explicitly state (a) the causal hypothesis under investigation in the case you’re
describing, (b) the expectation one can justifiably draw from the hypothesis, (c) some set of
observations relevant to evaluate the hypothesis, (d) the causal background, (e) one plausible
confounding factor, and (f) one method for causal inference one could justifiably employ in this case.
Finally, give one reason why/how the causal inference in the case you described can guide one
specific policy intervention, and give one reason why/how that causal inference is of little use in Comment [MC1]: Good – simplify it
policy making. though, focusing on bankruptcy or
underperformance, and specifying the
business (eg, amazon, nextflix, zoom, etc
did not underperfom...).
A) The causal statement I decided to focus on in this text is “The coronavirus is the main cause Also, single out the two variables you are
considering, eg X = Coronavirus (yes, no)
for businesses to underperform or go bankrupt in 2020”.
and Y = average revenue in biz of type B
Generally speaking, numerous economists predicted 2020 to be a prosperous year. This was (?). This will help better understand what
not the case since a great number of businesses in a lot of different industries experienced exactly the hyp is
financial distress. This happened at the same time as certain COVID-19 measures – such as Comment [MC2]: What type?
lockdowns which led to many employees being unable to go to work or even lose their job - Comment [MC3]: So is this a more
plausible proximate cause?
which were taken by various governments all around the world. Thus, the coronavirus and
Comment [MC4]: Yes, there is at least
businesses performing bad are correlated. an association
However, a correlation between two variables does not guarantee causation. In our case,
Comment [MC5]: Like what?
just because the virus spread more and more until it was basically all everywhere in the
Comment [MC6]: Okay par. rather than
world and at the same time businesses getting into financial trouble increased, doesn’t spending too much time on the
necessarily mean that one caused the other. It is possible that a confounding variable is the background, you could have defined the 2
vars under consideration
reason for the economic downturn.
Comment [MC7]: The first set of
B) There are some expectations which can be drawn from the hypothesis that the coronavirus considerations concerns a cue to causal
causes businesses to underperform or to go bankrupt. Firstly, spatiotemporal contiguity can relation (ie generally, causes and effects
are close in time and place --- though with
be viewed as a guide to causation. A correlation in time or in location can imply a causal economic phenomena this is less
relationship. Thus, we expect a certain spatiotemporal association between the two prominent); the second set of
considerations is about one way of
variables coronavirus and business facing financial distress. As already stated before, it is understanding what causality is.
obvious that these two variables are statistically dependent. We can also see that they
Best if you spelled out one
happened close in time because the financial distress succeeded the COVID-19 measures prediction/expectation from that hyp. At
immediately. Regarding the location, we can also not observe a “spatial gap” because the least, the reference to expectation was
meant that way (in line with the three
virus is basically everywhere and businesses did not underperform only in a specific region, ingredients of our recipes for science.).
but instead all around the world.
Comment [MC8]: What kind of
Secondly, we can expect something called “difference-making”. This means that when the observation is that?
happening of one event makes a difference to the occurrence of a second incident, it is Comment [MC9]: I want to hear... an
possible that the second event is a consequence of the first one. I assume that the answer to what’s been asked. Here is an
analogy.
emergence of the coronavirus made a difference to the global economy. I believe that if the
coronavirus and its measures had not happened, then a lot of businesses would not have Causal hyp: Eating pizza once a week
causes immunity to flu
underperformed or declared bankruptcy. Expectation: Populations who regularly eat
C) An observation relevant to evaluate the hypothesis is to see a dependence between the two pizza are less likely to catch the flu
variables corona virus and businesses facing financial distress. This was already discussed in Observation: I collect data from
part a and b. I am actually really not sure what you want to hear for this section and what I supermarkets and restaurants about the
rate of pizza eating in different countries in
should write… one year, and from healthcaredatasets
D) Now to the causal background: According to the textbook, a causal background of two about the rate of flu infections in the same
period. Comparing these observed data,
events can be defined as “all the other factors that actually do, or in principle might, causally one will be able to check whether and to
influence these two events, thereby also potentially affecting the causal relationship between what extent the expectation I drew from
my hyp is supported
