The word correlation seems to have magical powers till you enter a statistics class. The class rubbishes the claim “correlation implies causation” and with it dies all the magic associated with the word. So whenever you read the word correlation, your reactions are mixed.
This post by Matt Roberts Sklar of BoE Blog Bank Underground shows 250 years of correlation between bonds and equities. Interestingly, the correlation was positive till 1990s and has been negative ever since:
For most of the 18th-20th centuries, government bonds usually behaved like a risky asset. When equity prices fell, bond yields rose, i.e. bond and equity returns were positively correlated (bond prices move inversely to yields). But since the mid-2000s, bond and equity returns have been negatively correlated, i.e. bonds became a hedge for risk. Before this, the last time this correlation was near zero for a prolonged period was the long depression in the late 19th century.
Source: Thomas and Dimsdale (2016) and author calculations.
Line shows ten year trailing correlation of monthly returns.
The change in the bond-equity correlation since the mid-2000s partly reflects investors being less worried about inflation risks. As well as demand-type shocks being more prevalent than supply-type shocks, the introduction of credible inflation targeting has helped anchor inflation expectations and reduced the likelihood of high inflation risks. Investors may also have become more focussed on bad states of the world.
At the same time, there has been a structural increase in demand for ‘safe assets’, with more investors demanding safe government bonds for reasons unrelated to their expected cashflows. This has been exacerbated during and since the financial crisis, with deterioration in risk sentiment leading to episodic ‘flight to safety’. And the addition of QE and forward guidance to the monetary policy toolbox may mean long-term bonds react differently to previously.
The long historical perspective is interesting.
We usually think of stock and bond prices correlations like this. Say inflation goes up. Markets then expect central bank to increase rates and as a result bond yields go up and prices fall. In equity markets, the expectation is that valuation of firms will be lower in future down due to higher interest rates. So, the equity prices also decline. Thus, we see a positive correlation in the two assets.
But this monetary policy thing is recent development in 250 year history even if central banks are not. After all BoE came up in 1694, much before this data which is from 1750s. So to see this positive correlation even back in 1750s is interesting. Even more as during those times inflation was relatively stable due to commodity/gold standard. Prices hardly changed and inflation was hardly a risk. The shocks then mist be more around due to crop failures, sudden rise in price of gold/silver , banking panics etc.
So say there is a banking panic this would have led both bond and stock investors shunning both and prices declining.
I would imagine this correlation has changed mainly due to increased central bank adventurism. In case of a shock people know central banks will infuse liquidity mainly via buying more bonds. Thus bonds are never really out of favor and only equities suffer.
There is a lot of interesting stuff in this one graph..