In this blog, we regularly revisit our prediction of the rate of unemployment in Italy, which had been made in our 2008 (The previous revision was in 2017.) Five years after this publication, we found that the accuracy of prediction was excellent. We decided that our model works well. Since the model has a natural 11-year horizon, we were able to check our original (2008!) prediction for 2013 and 2016 and now for 2019 using new estimates of the unemployment rate in Italy (here we use the OECD database). According to the OECD, the unemployment rate in 2019 was 9.95%. For 2017, the rate was 11.2%.
Our model of unemployment as a function of the change in labour force predicts two pivot points in the unemployment rate – in 2008 and 2014. We introduced the model of unemployment in Italy in 2008 with data available only for 2006. The rate of unemployment was near its bottom at the level of 6%. The model predicted long-term growth in the rate of unemployment to the level of 11% in 2013-2014. The next pivot point is expected in 2023 and the rate will start to grow again.
The overall agreement between the measured and predicted unemployment estimates in Italy validates our concept, which states that there exists a long-term equilibrium link between unemployment, ut, and the rate of change of labour force, lt=dLF/LFdt. Italy is a unique economy to validate this link because the time lag of unemployment behind lt is eleven (!) years. The estimation method is standard – we seek for the best overall fit between observed and predicted curves by the LSQR method. The best-fit equation for the original data obtained from the national account is as follows:
ut = 5.0lt-11 + 0.07 (1)
As mentioned above, the lead of lt is eleven years. This defines the rate of unemployment many years ahead of the current change in the labour force. When the OECD data are used, the model has different coefficients:
ut = 3.0lt-11 + 0.09 (2)
The difference between national and international estimates of the labour force and unemployment is thoroughly discussed in our papers. We have also reported the change in definitions of GDP and CPI in several posts published this December.
Figure 1 repeats the picture from the 2017 post and displays the observed unemployment curve and that predicted using the rate of labour force change 11 years ago and equations (1). Since the estimates of the labour force in Italy are very noisy we have smoothed the annual predicted curve with MA(5). All in all, the predictive power of the model is excellent and timely fits major peaks and troughs after 1988. The period between 2006 and 2016 was predicted almost exactly.
In Figure 2 we add the most recent period and use the most recent version of the OECD data. The model (2) is presented in its annual and MA(5) versions. The prediction still works well and the next pivot point to increasing unemployment in Italy is expected in 2023.
The fit between predictions and observations is the best validation of any quantitative model. We do not know any other macroeconomic model capable to describe such dramatic turns many years ahead. The evolution of the rate of unemployment in Italy is completely defined ten years ahead. Since the linear coefficient in (2) is positive one needs to reduce the growth in the labour force in order to decrease the rate of unemployment.
Figure 1. Borrowed from the 2017 post. The observed and predicted rate of unemployment in Italy.