Historical data on IB short lending fee and share availability

Research
1

Hi,

Does anyone have historical data on IB short lending fee and share availability?

I’d like to analyze historical data to see what kind of lending fees my short sim would have experienced if I had actually traded it. I also want to analyze what happens if I skip trades for stocks that have very large fees.

I know you can download the current fees from IB’s FTP site (ftp3.interactivebrokers.com, username shortstock). However, there is no historical data on the FTP. IBorrowDesk.com has data going back one year, but it’s not point in time and data quality is questionable (I found several issues with it).

I’ve started to mirror IB’s FTP site every 15 minutes so I can build the historical dataset from now on. But I was wondering if any of you has been doing that already and is willing to share the data with me. I’m primarily interested in the ‘usa.txt’ file.

Thanks,
Peter

2

I have data going back to 2017-04-18. Downloaded between 4a and 7a each day.

P123 limits attachments to 20MB, so here’s a link: [url=https://gofile.io/?c=gRd3kH]https://gofile.io/?c=gRd3kH[/url]

3

That’s fantastic! Exactly what I was looking for.

Thanks for taking the time to upload and share it!

4

Hi,

I’m looking for the same historical data on IB short lending fee and share availability. The link above does not work anymore. Does anyone have this data?

Kriss

5

pvdb shared the data with me. A million thanks to him and SUpirate1081 for that!

6

I’ve also just started scraping the ftp site and can’t find the historical data anywhere. Anyone willing to share?
Thanks.

7

Here’s an updated link that should stay up: http://potential-investments.com/ib_shorting.zip

8

Many thanks! Appreciate it :slight_smile:

9

Has anyone got a good way to link the shorting data with P123 tickers? I'm running into trouble for tickers that have been de-listed because P123 puts the ^YY on the end.

Also any chance there is an updated file that includes 2021 to present?

10

Is there any chance P123 could add the short lending fee from FactSet? Apparently they have "financing rates data" through a provider called S3 so shouldn't be too hard?

There is a lots of signal in short lending fees as far as I can tell. Even more so than short interest, because short interest might be limited by how many shares are available to short. On the other hand, short lending fees truly reflect supply and demand to short the stock.

1 Like
11

Short interest / Institutional ownership is a proxy for lending fees

12

Respectfully, no.

1 Like
13

I think P123 ought to include some kind of borrow in the data set considering the types of analysis provided. We need a number, a cost, separately from how shorted the stock is.

1 Like
14

You just need to respect the paper:

Screenshot_20241018-070945
Screenshot_20241018-0709451440×400 143 KB

Yes, SIR/IO is a proxy for lending fee

22

OK I am sorry - I didn't mean to strike a nerve. Thanks for your input but I would still like actual short borrow fee data.

I'll now restate the original question after all this clutter, would it be possible for FactSet's "financing rates" to be added to the P123 database?

It looks like it is available through their partner S3.

23

I just checked 3 years of data that I had on hand, and it didn't look like a good proxy at all for IBKR borrow fees:

image

I didn't know FactSet had a financing rate in its dataset. Compelling for sure, and it seems like the path to p123 integration would be simple. But I would be over the moon if we could get IBKR borrow fees and available shares integrated into p123.

24

I've looked into the IBKR route and the problem is that they only keep historical data for tickers that are currently listed. Going back a year or two or even three would be fine but then you start having big gaps in the data for all the tickers that have since been delisted...

25

You forgot the "/IO" part

1 Like
26

Would that improve the R^2 from -0.02?

/sarc

1 Like
27

Yes, because high loan fees are largely due to low supply associated with low institutional ownership.

Edit: And the correlation between borrowing costs and SIR mentioned in the literature is 0.2 instead of -0.02