golfCaddy wrote: ↑
Tue Jun 12, 2018 10:58 am
It would be interesting if stock market data did go back 400 years, but I'm suspicious this is the financial equivalent of archeology. There appear to be substantial quality issues with US market data from the 1800s. If we can't trust the US data from the 1800s, can we trust the UK and Dutch data from the 1600s and 1700s?
It's a fair point that stock return data from the 1600s and 1700s is financial archaeology. According to the research, there were at most nine securities with recoverable trading histories for this period — two of which were Dutch (Dutch East India Co. and Dutch West India Co.) and seven of which were English (Bank of England, English East India Co., London Assurance Co., Million Bank, Royal African Co., Royal Exchange Assurance Co., and the South Sea Company).
Even with this very limited data set, though, it's interesting that real returns were in line with modern stock returns.
For bond returns, the deep history was pieced together from the best data available over the centuries (chart below), starting with northern Italy city markets in the 1300s, to Spain in the 1500s, then to Amsterdam in the 1600s, before moving to the U.K. and U.S. markets after 1700. If anything, these returns reflect a decline in the "societal risk premium" over time.
Note: Chart shows nominal bond returns, as opposed to real returns in the OP.
Certainly this deep history return data cannot compare in accuracy or completeness with modern data, but it's the best that's survived over the centuries — and it's not entirely without a few relevant lessons for modern times, in my view.