BBQ add-in

[dakila]

This thread is about the BBQ add-in that implements the Bry-Boschan (NBER) Business Cycle Dating Algorithm modified by Harding and Pagan for quarterly data

[terrya]

It appears that this isn't linking properly (as the BN filter did initially).

[dakila]

Yes, that link is not working. Gareth, please fix it.

[EViews Gareth]

Works for me - try refreshing.

[dakila]

Thanks it's working now.

[terrya]

Hi

With the following file when I run the add-in, the series GDPESA crashes resulting in the syntax error message. What's going wrong? The other 2 series give the right dating. (As an aside, the R package can evaluate GDPESA without problems.)

RGDPSANSA.xls

(34.5 KiB) Downloaded 1042 times

[terrya]

Further to this. Another data series, starting 1955Q1, wouldn't evaluate and the error message was "rows undefined". There's no problem with the data file when I use either R or Winrats 9.2. The recession periods generated in R and Winrats agree closely and are plausible, given the known events.

[dakila]

I fixed the bugs. Please update it.

[terrya]

Thanks will try late

[terrya]

Installed add-in as suggested. Definitely working. Thanks.

[joff]

Hi. I have tested the add-in using US GDP data from 1947q1 to 2017q3 over the whole sample the add-in works fine. However, when I play with the sample lenght I get two error messages (replication files are attached):

- Non numeric argument (e.g. for sample 1985q1 2017q3)
- Division by zero (e.g. for sample 1990q1 2017q3)

Since the program is encrypted could you please check for me what is causing these errors? I don't see the theoretical reason why the BB algorithm would not work in these cases.

Thank you in advance.

[dakila]

Hi, I fixed the bugs. Please update it.

[dakila]

Hi there,
Please update the add-in. Now it can handle series with NA correctly.