Question

Reconsidering the Baumol results. J. Bradford DeLong (1988), in

a comment on Baumol’s convergence result for the industrialized

countries over the last century, pointed out that the result could be

driven by the procedure through which the countries were selected.

In particular, DeLong noted two things. First, only countries that

were rich at the end of the sample (i.e., in the 1980s) were included.

Second, several countries not included, such as Argentina, were

richer than Japan in 1870. Use these points to criticize and discuss

the Baumol results. Do these criticisms apply to the results for the

OECD? For the world?

Answer 1

Here the result of Baumol's study where he tried to establish the long-run convergence of the per capita incomes of the countries as predicted by the Solow model. he considered only sixteen countries data which were available in Madison's database. He took data of these countries from 1870 to 1979: Japan, Finland, Sweden, Norway, Germany, Italy, Austria, France, Canada, Denmark, the United States, the Netherlands, Switzerland, Belgium, the United Kingdom, and Australia

He found out that

1. the countries with lower per capita income in the 1870s showed a higher growth rate.

2. the countries, when they got richer per capita income-wise their growth, slowed down.

Hence he is supposedly giving proof that Solow's neoclassical growth model indeed predicted correctly.

but as Delong pointed out Baumol made some serious statistical errors:

Selection Bias: The countries which had their data available in 1870 has one common characteristic: that they were rich countries even in 1870 and hence they had the technology and knowledge advancement to collect and store the data. hence the results he provided cannot be applied to the other countries of the world because the selection of the countries was not random. This is where Delong argued why Argentina was not included, why Japan. because Japan had data means they were rich beforehand.

Delong also made these points:

a. DeLong added other seven countries to the list and then he ran a new regression to find out that the goodness of fit of the model is not good. that means the log per capita income of 1870 does a poor job explaining the log per capita growth of the countries. hence Baumol's result became an unclear picture.

b. DeLong also added that the measurement errors are likely to be high in 1870 than in 1979. hence the data was full of errors and hence the result derived is ambiguous.

these criticism don't apply to the OECD countries. because Baumol took the sample of only rich countries and OECD countries are/were very rich Per capita income-wise.

but these criticisms apply to the world as a whole because the world has much more numbers of countries where there are a lot of poor countries and many countries don't have data available. hence these two criticisms apply to this context.