Cross-country comparison: labor force participation in Canada and Italy

One of the most important requirements to a sound macroeconomic model is the capability to explain the difference in evolution of modelled parameters across developed countries.  For example, a consistent model of the rate of participation in labour force, LFP, has to describe the striking difference observed in the long-term behaviour of LFP in Canada and Italy. Figures 1 and 2 depict the measured (open circless), as provided by the BLS: http://www.bls.gov/data/, and predicted LFP. The latter is obtained from the model linking LFP to real GDP per capita only [1]. The GDP estimates are taken from the Conference Board data base (at GK PPPs).

Our model shows an exceptional predictive power for both countries. This accurate prediction is obtained despite the measured LFP in Canada has been growing since 1960 and that in Italy has been on decline.  Moreover, even short-term deviations from the overall trend are well predicted in time and amplitude. In Figures 1 and 2 we added two new measurements made in 2008 and 2009 to the original curves published in [1]. 

One can conclude that the model does not contradict actual measurements in Sweden, Canada, and Italy.  We are going to extend the set of countries and the duration of relevant time series.

Figure 1. Measured and predicted LFP in Canada.                  

Figure 2. Measured and predicted LFP in Italy. 

References

1. Ivan O. KITOV, 2008. "The Driving Force of Labor Force Participation in Developed Countries," Journal of Applied Economic Sciences, Spiru Haret University, Faculty of Financial Management and Accounting Craiova, vol. 3(3(5)_Fall), pages 203-222.

Labor force participation in Sweden

The labor productivity model discussed in the previous post is based on the concept linking labor force participation rate, LFP, to real GDP per capita [1]. This is a primary model, which explains the dynamics and the long-term behavior of labor force level in developed countries. As before, the LFP model is extremely parsimonious and uses only one (!) defining parameter to explain all variations in the observed behaviou of labor force in developed countries. As a consequence, one needs no other macro- or micro-economic variable to explain the portion of labor in total population.  

In this post, we do not formally introduce the quantitative model since it is available in the paper and monograph. Our purpose is to extend the previous data set by two years (2008 and 2009) since new observations are now available. This is in line with our validation strategy – to test all models with new data.

Figure 3.13 is borrowed from our monograph and illustrates the predictive power of the model for Sweden. The agreement between the original LFP estimates (open circles) and those predicted by the model is excellent in timing and amplitude. Considering the fact that only one defining variable is used the prediction suggests the presence of long-term on-to-one link between LFP and real GDP. (More examples in the paper and monograph.)

Figure 1 extends the original data set by two estimates (real GDP per capita reported by the Conference Board). The agreement is also excellent.  This observation evidences in favor of our model.

We will continue reporting the accuracy of LFP predictions for Sweden and other developed countries.

Figure 3.13. Observed and predicted growth rate of LFP in Sweden: N(1959)=100000, A₂=$310 (1990 U.S. dollars), B=2.2∙10⁶, C=-0.0465, T=0. Lower panel depicts the original LFP, changing in the range from 67% in 1990 to 62.5 % in 1998, and the predicted LFP.

Figure 1. Same as in Figure 3.13, but extended with data in 2008 and 2009. The original LFP series is  reported by the BLS.

References

1. Ivan O. KITOV, 2008. "The Driving Force of Labor Force Participation in Developed Countries," Journal of Applied Economic Sciences, Spiru Haret University, Faculty of Financial Management and Accounting Craiova, vol. 3(3(5)_Fall), pages 203-222.

Labor productivity in Canada

Two years ago we published two papers [1,2] which introduced a new macroeconomic model explaining the evolution of labor productivity in developed countries. The model is absolutely parsimonious and uses only one measured macroeconomic variable as the driving force of the productivity growth – real GDP per capita. Figure 3.22 is borrowed from our monograph “mechanomics. Economics as Classical Mechanics” and illustrates the predictive power of the model as applied to Canada. (Due to extremely high volatility of productivity measurements, the measured data set is represented by its 5-year moving average, MA(5)).  All coefficients in the model for Canada were obtained empirically, as explained in the monograph.

Considering the simplicity of the model and the accuracy of data on real GDP and productivity, the prediction of the time history in Canada is excellent. (We would be grateful if the reader could provide us with a reference to a model which gives better predictions.) It is also important that the prediction covers the whole period since 1960 with one deterministic link without any structural breaks. The latter is the inevitable and crucial element of any explanation of productivity in developed countries. Moreover, all mainstream macroeconomic models (e.g. DGSE) are using the notion of shocks to productivity as a central phenomenon explaining all bigger deviations in the rate of real economic growth. This implies that productivity must define real GDP. This assumption contradicts observations, as our model demonstrates – the change in labor force productivity lags by two (!) years behind the defining change in real GDP. Therefore, productivity is not a proactive macroeconomic variable.

Figure 3.22. Observed and predicted productivity in Canada: N(1959)=270000, A₂=$300, B=-3200000, C=0.108; R²=0.8.

Since the data set was limited by 2007, one can test the predictive power of the model using new data and extend the forecasting horizon. As before, we use the data set published by the Conference Board [3].  Figure 1 shows that our prediction for 2008 and 2009 was accurate. In the near future, one can expect a significant growth in labor productivity in Canada.

For further validation of the model, we are going to revisit our predictions for other developed countries.

Figure 1. Same as in Figure 3.22 extended by measurements in 2008 and 2009.

References

1. Kitov, I., Kitov, O., (2008). The driving force of labor productivity, MPRA Paper 9069, University Library of Munich, Germany, http://ideas.repec.org/p/pra/mprapa/9069.html

http://mpra.ub.uni-muenchen.de/9069/01/MPRA_paper_9069.pdf

2. Kitov, I., Kitov, O., (2009). Modelling and predicting labor force productivity, MPRA Paper 15152, University Library of Munich, Germany, http://mpra.ub.uni-muenchen.de/15152/01/MPRA_paper_15152.pdf

3. Conference Board. (2010). Total Economy Database, January 2010. http://www.conference-board.org/data/economydatabase/

CPI and core CPI

Following our previous post on the difference between the headline CPI and GDP deflator we have to revisit the difference between the overall CPI and core CPI (the headline CPI less food and energy). This difference also demonstrates severe changes in definitions of all involved variables and reveals a break in these time series, which harms the compatibility of CPI measurements before and after 1979.

Two years ago we published a paper on the presence of long-term sustainable trends in the differences between various components of the CPI in the USA. We started with the difference between the core CPI (i.e. CPI less food and energy) and the overall CPI. Figure 1 is similar to Figures 1 and 2 in the paper.

Figure 1. Linear regression of the difference between the core CPI and CPI for the period from 1981 to 1999 (R²= 0.96 the slope is 0.67) and linear regression of the difference between the core CPI and CPI between 2002 and 2009 (R²=0.91, and the slope is -1.59).  

We also suggested in this and later papers on the sustainable trends in the CPI and PPI (see here) that the negative trend after 2002 should reach some bottom point and turn to a positive trend. It was also mentioned that such processes in the past had been accompanied by an elevated volatility in the difference, i.e. high amplitude fluctuations. All predictions were actually observed. Figure 2 updates Figure 1 with data available in November 2010.

Figure 2. A new positive trend has been emerging since 2010. 

Therefore, we confirm our previous predictions and expect the new positive trend in Figure 2 shows. This trend repeats the trend observed between 1987 and 1999 rather than the mirror reflection of the previous negative trend between 2002 and 2009, as was suggested before. Thus, the price indices of food and energy will not be falling too fast relative to the core CPI, but this period will likely last more than 10 years. We will keep posting on the difference

Is real GDP correct?

This is an instructive story about the metrology of macroeconomic measurements. From the point of view of hard sciences, the economics profession does a great counterproductive work in order to hide actual links between macroeconomics variables. Among the most effective tools of this clandestine operation is the random change in definitions of these macroeconomics variables without mentioning it during statistical analysis.   We have devoted enough time to reveal and recover many trivial cases in our book “mecħanomics. Economic as Classical Mechanics”, but the list is too far to be closed yet.

It is well known that there is no such macroeconomic measurable parameter as real GDP (see Concepts and Methods of the U.S. NIPA for details). There are two actually measured variables: nominal GDP and GDP deflator (price index). Real GDP is estimated using nominal GDP less the change in prices. (General public usually thinks that there are real and nominal GDP estimated and the difference is called price inflation. Wrong.) The latter is not easy to calculate or even evaluate.  It is so much sophisticated problem that before 1980, there was no practical difference between the consumer price index (CPI) and the GDP deflator in the U.S., as Figure 1 demonstrates. Effectively, the curves in the Figure split in 1980. There is no direct statement about the reasons of the change in definitions in the aforementioned conceptual document, but we might guess that this is likely related to the introduction of new methodology to evaluate price inflation. 

Thus, before 1980 the CPI was used as an estimate of price inflation. Since 1980, GDP deflator has been used. The difference between these two variables can not be neglected: the cumulative change in inflation between 1980 and 2009 is 20 points – the Figure shows cumulative change in inflation rate since 1929. Does that mean that when applied to the estimates before 1980, the concept of dGDP would result in even bigger change in real GDP estimates?

All in all, the notion of real GDP is a virtual one and is highly biased by the change in its definition in 1979.  One has to be very careful when using real GDP estimates in economic analysis. Do not trust the BEA before you check the comparability of their estimates through time.

Figure 1. Cumulative  rate (the sum of annual inflation rates, what is different from inflation index) of inflation since 1929, as described by the CPI and dGDP.

The price index of motor fuel

Couple years ago we introduced a concept of deterministic and sustainable trends in the differences of consumer (and later on - producer) price indices [1,2]. One of the best examples was the difference between the overall (often called “headline”) CPI and the index of motor oil [3], both indices were seasonally adjusted ones. A year ago we revisited this difference [4] and found our predictions in good agreement with observations. Essentially, this difference approached the new trend and we expected the following evolution along this new trend, which assumes that the price index of motor fuel will decrease relative to the CPI.

The model implies that the difference between the overall CPI (same for the PPI), CPI (PPI), and a given individual price index iCPI (iPPI), can be described by a linear time function over time intervals of several years:

CPI(t) – iCPI(t) = A + Bt (1)

, where A and B are the regression coefficients, and t is the elapsed time. Therefore, the “distance” between the CPI and the studied index is a linear function of time, with a positive or negative slope B. Free term A compensates the difference related to the start levels for a given year. For example, the index of communication was started from the level of 100 in December 1997 when the overall CPI was already at the level of 161.8 (base period 1982-84 =100).

This post displays the evolution of the difference in 2010 (see Figure 1). All in all, the prediction was good enough: the earlier positive deviation from the trend is now compensated by a negative one. Currently, the difference is below the trend and we expect it to reach the trend in the beginning of 2011. This recovery should be accompanied by a drop in the price index of motor fuel and likely in the index of crude oil.

Figure 1. The difference between the headline CPI and the index for motor fuel. Solid diamonds represent the prediction given in March 2009 through December 2009 [1]. The total increase in the difference is +60 units of index or +35%: from 173 in March to 233 in December. Dashed line represents the new trend, which is a mirror reflection to that between 2001 and 2008 shown by solid black line. In 2010, the difference has been fluctuating around the trend and thus should return to the trend in the beginning of 2011.

References

1. Kitov, I., Kitov, O. (2008). Long-Term Linear Trends In Consumer Price Indices, Journal of Applied Economic Sciences, Spiru Haret University, Faculty of Financial Management and Accounting Craiova, vol. III(2(4)_Summ), pp. 101-112.

2. Kitov, I., Kitov, O. (2009). Sustainable trends in producer price indices, Journal of Applied Research in Finance, vol. I(1(1)_ Summ), pp. 43-51.

3. Kitov, I., Kitov, O. (2009). A fair price for motor fuel in the United States, MPRA Paper 15039, University Library of Munich, Germany

4. Kitov, I.. Kitov, O. (2010). "Crude oil and motor fuel: Fair price revisited," Quantitative Finance Papers 1005.0051, arXiv.org.

Bernanke and deflation

Deflation is the economic topic where the current  Chairman of the United States Federal Reserve Shalom Bernanke has got his nickname. We have been following the evolution of inflation in the USA for many years since predicted (2005)  that a deflationary period should start in 2012.  So, it is instructive to  compare the tools proposed by the Fed's chairman and actual situation. The most recent statement from the FRB on deflation was in the speech  on October 15.

... The significant moderation in price increases has been widespread across many categories of spending, as is evident from various measures that exclude the most extreme price movements in each period. For example, the so-called trimmed mean consumer price index (CPI) has risen by only 0.9 percent over the past 12 months, and a related measure, the median CPI, has increased by only 0.5 percent over the same period.2



... With long-run inflation expectations stable and with substantial resource slack continuing to restrain cost pressures, it seems likely that inflation trends will remain subdued for some time.

The longer-run inflation projections in the SEP indicate that FOMC participants generally judge the mandate-consistent inflation rate to be about 2 percent or a bit below. In contrast, as I noted earlier, recent readings on underlying inflation have been approximately 1 percent. Thus, in effect, inflation is running at rates that are too low relative to the levels that the Committee judges to be most consistent with the Federal Reserve's dual mandate in the longer run. In particular, at current rates of inflation, the constraint imposed by the zero lower bound on nominal interest rates is too tight (the short-term real interest rate is too high, given the state of the economy), and the risk of deflation is higher than desirable. Given that monetary policy works with a lag, the more relevant question is whether this situation is forecast to continue. In light of the recent decline in inflation, the degree of slack in the economy, and the relative stability of inflation expectations, it is reasonable to forecast that underlying inflation--setting aside the inevitable short-run volatility--will be less than the mandate-consistent inflation rate for some time. Of course, forecasts of inflation, as of other key economic variables, are uncertain and must be regularly updated with the arrival of new information.

and finally

Given the Committee's objectives, there would appear--all else being equal--to be a case for further action. However, as I indicated earlier, one of the implications of a low-inflation environment is that policy is more likely to be constrained by the fact that nominal interest rates cannot be reduced below zero. Indeed, the Federal Reserve reduced its target for the federal funds rate to a range of 0 to 25 basis points almost two years ago, in December 2008. Further policy accommodation is certainly possible even with the overnight interest rate at zero, but nonconventional policies have costs and limitations that must be taken into account in judging whether and how aggressively they should be used ...

So, the problem of approacing price deflation now is a big one. To implememnt a helicopter technology is not so easy.