Clark (1985)

Article
Has bibtex key
Has article title
Has author	Clark
Has year	1985
In journal
In volume
In number
Has pages
Has publisher
© edegan.com, 2016

Theory Questions:

What is the research question?

The article attempts to answer the question: "Why were New England cotton textile workers so much more productive than similar workers in the rest of the world?"

What is the author's hypothesis?

The author believes that local culture explains the difference and that differences in technology, management, worker training and inherent worker ability do not satisfy as explanations.

How does the author test the hypothesis?

The author never directly tests the hypothesis, but goes about "proving" his case by examining and disproving a thorough set of alternative hypotheses.

In addition, the author shows that there was a very large difference in the average amount of machines each worker attended across countries.

How does the author rule out alternative hypotheses?

The author addresses a number of alternative hypotheses about why New England was so productive:

• Are machines more efficient across countries? Table 5 shows that they aren't.
• Were some countries using more labor-per-machine to save on cotton costs? IE: Buy cheaper cotton, use more labor. Authors say this is not happening for five reasons listed on page 156. Among those reasons: (1) Worse cotton would have reduced overall product quality, (2) Even processes in the mill having nothing to do with cotton quality were overstaffed, (3) The countries and factories with an overstaffing problem were not the ones using cheaper yarn.
• Technological differences: Basically everyone used the same machines, often bought from the same source.
• Amount of experience of workers: Authors note that many countries hired temporary workers without much impact on productivity, which suggests that returns to experience are small. They also note that there was no effort to try to codify the learnings of older, more experienced workers.
• Differences in inherit capability of workers: The authors note that little strength or intelligence was required to perform the tasks, so this should not have made a difference.

Related to the final point, the authors also note that in New England, a lot of the workers were immigrants from Ireland, where productivity was a lot less. If "inherit ability" was an issue, the Irish immigrants should have been less productive. Verbal arguments about selection offered on pg 167.

How might these tests be run if one had quantitative evidence?

One could test the hypotheses above the technology, worker experience and etc had no effect on worker productivity by finding a source of random, exogenous variation.

In particular, one might find be able to test this the effect of experience. The authors report that in some countries -- both temporary as well permanent staff were employed during different periods of the sample period. This might create an opportunity for a differences-in-differences approach to see how productive less experienced employees are.

What problems might arise in this quantitative analysis?

The biggest problem is that local culture -- the author's primary driving force behind differences in outcomes -- is hard to distribute randomly.