Hens Steehouwer’s ‘stylised facts’ are the building blocks for the kind of scenario analysis that pension funds need for their asset liability models, and for everybody else interested in the behaviour of macroeconomic variables.
They are defined as “observed phenomena which are found to be robust with respect to the historical time period and/or the country of origin for which they are observed.”
Analysing time-series by means of a frequency domain approach (see main article) can increase our understanding of the behaviour of macroeconomic variables. As an example of a stylised fact consider the inflation hedging capacities of equities for which it is not difficult to imagine the importance for long term asset allocation choices for pension funds.
Steehouwer points out in his extensive Phd study ‘Macroeconomic Scenarios and Reality’ that the economic intuition is that, unlike nominal bonds, equities are a claim on real assets and that their value should therefore in some way move in line with (expected) changes in the general price level.
Economists have extensively tested this hypothesis and found that the evidence is ambiguous. Eugene Fama and William Schwert for example were one of the first to find that US stock prices are negatively correlated to current and lagged inflation. Other research has lent weight to the idea that equities are, in fact, a poor hedge against inflation.
Bruno Solnik summarised the case for a negative correlation by saying that “there is a consistent lack of positive association between stock returns and inflation covering several countries”.
However, Steehouwer notes that most of the reported negative correlations relate to the short term. The long term results were rather different. Other researchers who looked at two centuries of UK and US data have found a positive correlation between five year equity returns and inflation or, in general, when longer horizons were considered.
Steehouwer therefore decided to make a split between analysing the short and long term relation between equities and inflation. Using the frequency domain filtering approach (see main article) it is easy and natural to define the business cycle fluctuations as the short term fluctuations and the long wave and other long term swings as the long term fluctuations.
The results showed that the correlations were consistently negative in the short term but consistently positive in the long term. The short term results make sense in terms of the business cycle, he says, “We know that consumer prices lag the business cycle in the national product by approximately two to four years.”
“We also know however that equity prices lead the business cycle by approximately one year. It is then not hard to imagine that these very different phases imply a negative short term conventional correlation between equities and consumer prices.
“In short, the negative short term relation between equities and consumer prices is a direct consequence of their specific lead/lag relations in terms of the business cycle.”
The positive correlations confirm the indications in the literature of the existence of a positive long term correlation between equities and prices.
Steehouwer says that “the conclusion must be that, in the short run, equities are a poor hedge against inflation because of the substantial phase differences between equities and consumer prices in terms of business cycle fluctuations. In the long run equities do provide a good hedge against inflation as economic intuition tells us.”