Bond Yield Paradoxes: Gibson's Paradox and Shiller's Paradox

Today, there is a widely-held notion that government bond yields should reflect a long-term inflation expectation among other factors: such as “real risk-free rate”, “default premium”, “illiquidity premium”, “maturity premium”, etc.

Nevertheless, in history prior to the 1980s, there was a totally different widely-held consensus regarding bond yields. In the United Kingdom before the 1980s bond yields of the UK Console demonstrated a high correlation (77.7% during the period of 1737-1981) with the general price level (the log of price level), instead of inflation expectation (Chart 3.2.1). And it was known as “Gibson’s Paradox”. John Maynard Keynes characterised the historical relationship as:?

“one of the most completely established empirical facts in the whole field of quantitative economies.” (Keynes, 1965, vol 2, p198)

So, Keynes and other economic minds of his time used to see that bond yields are a reflection of current price level, but not inflation expectation. The notion is unconceivable today. But, Chart 3.2.1 below illustrates the notion of “Gibson’s Paradox”.

Nevertheless, after the life of Keynes, something happened to transform the relationship.?

A new paradox emerged in the new relationship between bond yields and inflation. It was articulated by Robert J. Shiller.

Shiller (2015, pp. 42-43) suggested a paradoxical relationship between inflation rates and bond yields, positing a revisionist notion that bond yields might better serve as a rear-view mirror to reflect past inflation experiences, rather than forward-looking inflation expectations.

Chart 3.2.2.a illustrates "Shiller’s Paradox" over UK Consols historical data for about three centuries. Along historical bond yields, Shiller projected two sorts of 10-year average inflation rates―one retrospective and the other prospective―and empirically illustrated that bond yields have a higher correlation with the retrospective inflation average than with the prospective alternative. (Shiller, 2015, pp. 42-43)

And Chart 3.2.2.b illustrates Shiller’s paradox using US data.

From the 1930s to the early 1960s, in the US case a material diversion occurred between these two metrics: the trailing 10Y inflation average and the bond yields. It is partly because of the government intervention in yield curve during WWII (1938-1945) for war finances.

During WWII, the US Treasury controlled yields to shape a stable rising slope in the yield curve.?The Federal Reserve Banks purchased the securities in order to achieve the fixed yield curve schedule. This guaranteed that bondholders would benefit from “riding the yield curve”: towards maturity, the bond price increases as scheduled. In this way, the US Treasury promoted investment in US war bonds. (Homer & Sylla, 2005, p. 354)

And as we can observe, these two metrics gradually converged thereafter.

Whether you agree with their arguments or not, there is one revelation from these empirical observations. Both of them invalidate human predictive power regarding future inflation.

If you are interested in historical relationships between inflation and bond yields, there are more issues to be discussed.

• Government’s yield management
• Geopolitical conflicts (Hegemonic wars)
• Changes in Inflation/Deflation Cycle
• Inflation volatility
• Non-monetary interactions: innovation, psychology of scarcity/abundance of supply, employment implications

If interested, please visit my old article in the following link: https://www.reversalpoint.com/bond-wave-mapping-2-price--inflation-cycles.html

References

• Keynes, J. M. (1965). A treatise on money: Vol. II, The applied theory of money. London: Macmillan & Co Ltd.
• Shiller, R. J. (2015). Irrational Exuberance (3rd Edition ed.). Princeton, New Jersey: Princeton University Press.

#prices #inflationexpectations #bondyields #investment