Problem in BEA data : order of magnitude


#1

Hi, we realized a problem with the nominal GDP data for the US.
We download the data in R via the following code:

data_gdp_us <- rdb('BEA', 'NIPA-T10105', mask = 'A191RC-Q')

Then we ask the value for 2018-Q4:

data_gdp_us$value[which(data_gdp_us$period == "2018-10-01")]

The answer is 20897804 which is roughly 21 x 10^6.
The column “dataset_name” tells us that the data are in billions USD (=10^9 in short scale which is the one used in the US):

data_gdp_us$dataset_name[which(data_gdp_us$period == "2018-10-01")]

Thus, we have 21 x 10^15 USD.
According to the BEA press release the 2018-Q4 nominal GDP in the US was 20.87 trillion, thus roughly 21 x 10^12 USD.
When consulting the NIPA table Table 1.1.5 directly at the BEA (it is in billions USD), the number 20 897.8 is shown for 2018-Q4, thus once again 21 x 10^12 USD.
Please let me know when this issue is solved so that we can change or codes accordingly.
Thanks a lot,
Julia


#2

Thank you Julia for reporting the problem. There is an inconsistency in the data provided but the BEA API.

I have open issue https://git.nomics.world/dbnomics-fetchers/management/issues/491 to find a solution to this problem. We will let you know when it is implemented.

Thanks for helping us improve DBnomics

Michel


#3

Hi Julia, I’m Bruno from the DBnomics team. We fixed this problem with BEA;

  • we added two dimensions: Metric and Unit that are computed from series-level information from BEA API, so this new information should be correct
  • we removed the unit given by BEA in dataset name as this is false in many cases (like the one you pointed out)

In the example you gave here, the NIPA-T10105/A191RC series have a dimension Metric with value Millions of current Dollars, which is correct.

Do not hesitate to let us informed if you find another problem !
Thanks