Josh Angrist: Are Machine Learning and Big Data Changing Econometrics?

♪ [music] ♪

- [Narrator] Are machine learning
and big data changing econometrics?

[gong ringing]

- [Josh] Well, machine learning
is certainly interesting and useful.

It's capturing people's attention,
and, in some cases,

it's leading to interesting new
methodological work.

I don't believe it's going
to have a big effect

on applied econometrics,

at least not
for the sort of questions

that I'm interested in.

For one thing,

a lot of the data sets that we use
to answer causal questions

are not particularly big,

so there isn't much scope
for big data methods.

A lot of the bigness seems to me
to come from commercial activity --

so research that's tied
to commercial activities,

fundamentally marketing research.

In some cases, I'm a little worried

that the "big" in big data
is kind of becoming a distraction.

This happens every time there's
a new set of tools on the scene.

Analysts get interested.

Of course young scholars,
graduate students,

are looking for a way
to make their mark.

For the most part,
the tools that I use,

and people who are interested

in the kind of questions
that we cover

in the Mastering Metrics
video series,

have been around for decades,

and I don't see that they're likely
to be changed much

by big data and machine learning.

You could say,
"Well, that's one man's opinion,"

and that's true,

but I don't see the application
that's important and useful

that depends on the use
of these new tools.

I can't think of any examples.

♪ [music] ♪

- [Narrator] Ready
to master econometrics?

Click here to embark
on an educational journey

with Josh Angrist,
a.k.a. "Master Joshway."

Or, if you'd like to watch more
from this interview series,

click here.

♪ [music] ♪