Regressing variables w/ different timeframe

[Anass Wajib]

Hello,

I'm conducting a study on the likelihood between R&D spending by country and innovation (patent applications as its proxy), I'm considering a panel data to run my regressions, as much as I'm trying to make that fair and comparable, I'm currently faced with this issue being my patent applications dataset starts from 1963 and ends on 1999 while the furthest R&D spending by country, I did found was that of OECD and it starts from 1983 to 2015.

I'm thinking by regressing these two it won't be at all accurate, and borderline misleading. Any idea to sort that out you think?

Thank you in advance.

[Carlo Lazzaro]

Anass:
this is a matter of data availability,
The only epirical regression that you can do includes data from 1983 to 1999.
Otherwise, you may want to simulate/impute the missing data according to time frame you decide to consider.