Tweet
It is worth to repeat the analysis done for ExxonMobile for another oil company (for more details about oil price and CPI see my blog http://inflationusa.blogspot.com/). ConocoPhillips is a good candidate, as mentioned by nero3.
In a sense, company name is irrelevant.
This is the approach that is important. One might try any company and foresee when the price will go up or down, except those companies which do not follow the link to the difference (core CPI-CPI). If you find one, please let me know and I’ll try to invent something else. I hope that the forecast at a 5-year horizon for stock prices is a valuable asset itself.
The link between the price for energy and consumer price index allows suggesting that stock prices related to energy might be driven by the difference between the core and headline CPI.
ConocoPhillips (COP) is aslo a typical representative of energy industry and the company with a major weight in S&P 500. We borrowed monthly prices from COP from Yahoo finance -
http://finance.yahoo.com/q/hp?s=COP&...=17&f=2009&g=m
Figure 3 displays two curves - historical (close) price for ConocoPhillips stocks (black line) and the difference between the core CPI and headline CPI between 1982 and 2009. The latter is scaled by the following function:
COP(t) =(-4)*(core CPI- CPI) + 90 (1)
where COP(t) is the predicted COP price which fits two curves between 2000 and 2008. There is no fit before 2000, however. At first glance, one might assume that the difference provides no information about the evolution of the price for COP. This is not a right assumption. Figure 1 shows that linear trend before 2000 was a positive one and after 2000 – negative. So, it is reasonable to suggest that the sign of slope in (1) should be changed to an opposite one: +4. Free term in (1) is another issue – it may change randomly. After reversing the sign and calibrating relevant amplitude and level between 1980 and 2001 by the function:
COP(t) = (+4)*(core CPI- CPI) + 0; from 1980 to 2000 (2)
we obtain a much better fit as depicted in Figure 4.
Now we can conclude that the difference between the core CPI and the headline CPI is a very good approximation of the evolution of the price of energy-related stocks (COP and XOM), except short periods of spike-like behavior. The spikes are definitely related to stock split in 1985 and 2005. Later on we will process COP price adjusted for dividents and stock splits.
Figure 5 shows the most recent period between 2000 and 2009. The fit between the observed and predicted (from the CPI) price is remarkable. Although, this is not a fundamental finding, because the CPI includes the price index for energy. To some extent, we correlate the price index for energy and energy stock price.
The fundamental findings are as follows:
1. During the short periods of transition between linear trends, as shown in Figure 3, the price for energy-related stocks (for example, COP) suffers a strong fall! ( Watch the next turn in the CPI trend!) It might be induced by an elevated volatility in the CPI associated with the transitions.
2. The dependence of the price for energy-related stocks on the difference between core CPI and CPI currently likely undergoes a transformation from a negative factor (-4) to a positive one (+4) (see Figure 1)
3. Because the difference (core CPI- CPI) will be positive during the next 5 to 10 years, COP stock price is likely to grow despite the price index for energy will be growing at lower rate than the core CPI (see Figure 2).
4. A new rally will likely start by the end of the current transition period.
Figure 1. The difference between core CPI and headline CPI as a function of time. One can distinguish three lengthy periods of quasi-linear behavior. Since the second half of 2008, the difference has been suffering a transition to a new linear trend with a positive slope.
Figure 2. The difference between the core CPI and the index for energy is similar to that between the core CPI and CPI in Figure 1.
Figure 3. Historical (close) price for COP (black line) and the difference between the core CPI and headline CPI between 1982 and 2009. The latter is scaled by the following function: -4*(core CPI- CPI) + 90, in order to fit two curves between 2001 and 2009. Two major spikes in the COP curve are related to stock splits in 1985 and 2005.
Figure 4. Historical (close) price for COP (black line) and the difference between the core CPI and headline CPI between 1970 and 2009. The latter is scaled by the following time dependent function (+4)*(core CPI-CPI) + 0, between 1980 and 2001; (-4)*(core CPI- CPI) + 90, between 2001 and 2009.
Figure 5. Comparison of the observed and predicted price for COP stock between 2000 and 2009
In a sense, company name is irrelevant.
This is the approach that is important. One might try any company and foresee when the price will go up or down, except those companies which do not follow the link to the difference (core CPI-CPI). If you find one, please let me know and I’ll try to invent something else. I hope that the forecast at a 5-year horizon for stock prices is a valuable asset itself.
The link between the price for energy and consumer price index allows suggesting that stock prices related to energy might be driven by the difference between the core and headline CPI.
ConocoPhillips (COP) is aslo a typical representative of energy industry and the company with a major weight in S&P 500. We borrowed monthly prices from COP from Yahoo finance -
http://finance.yahoo.com/q/hp?s=COP&...=17&f=2009&g=m
Figure 3 displays two curves - historical (close) price for ConocoPhillips stocks (black line) and the difference between the core CPI and headline CPI between 1982 and 2009. The latter is scaled by the following function:
COP(t) =(-4)*(core CPI- CPI) + 90 (1)
where COP(t) is the predicted COP price which fits two curves between 2000 and 2008. There is no fit before 2000, however. At first glance, one might assume that the difference provides no information about the evolution of the price for COP. This is not a right assumption. Figure 1 shows that linear trend before 2000 was a positive one and after 2000 – negative. So, it is reasonable to suggest that the sign of slope in (1) should be changed to an opposite one: +4. Free term in (1) is another issue – it may change randomly. After reversing the sign and calibrating relevant amplitude and level between 1980 and 2001 by the function:
COP(t) = (+4)*(core CPI- CPI) + 0; from 1980 to 2000 (2)
we obtain a much better fit as depicted in Figure 4.
Now we can conclude that the difference between the core CPI and the headline CPI is a very good approximation of the evolution of the price of energy-related stocks (COP and XOM), except short periods of spike-like behavior. The spikes are definitely related to stock split in 1985 and 2005. Later on we will process COP price adjusted for dividents and stock splits.
Figure 5 shows the most recent period between 2000 and 2009. The fit between the observed and predicted (from the CPI) price is remarkable. Although, this is not a fundamental finding, because the CPI includes the price index for energy. To some extent, we correlate the price index for energy and energy stock price.
The fundamental findings are as follows:
1. During the short periods of transition between linear trends, as shown in Figure 3, the price for energy-related stocks (for example, COP) suffers a strong fall! ( Watch the next turn in the CPI trend!) It might be induced by an elevated volatility in the CPI associated with the transitions.
2. The dependence of the price for energy-related stocks on the difference between core CPI and CPI currently likely undergoes a transformation from a negative factor (-4) to a positive one (+4) (see Figure 1)
3. Because the difference (core CPI- CPI) will be positive during the next 5 to 10 years, COP stock price is likely to grow despite the price index for energy will be growing at lower rate than the core CPI (see Figure 2).
4. A new rally will likely start by the end of the current transition period.
Figure 1. The difference between core CPI and headline CPI as a function of time. One can distinguish three lengthy periods of quasi-linear behavior. Since the second half of 2008, the difference has been suffering a transition to a new linear trend with a positive slope.
Figure 2. The difference between the core CPI and the index for energy is similar to that between the core CPI and CPI in Figure 1.
Figure 3. Historical (close) price for COP (black line) and the difference between the core CPI and headline CPI between 1982 and 2009. The latter is scaled by the following function: -4*(core CPI- CPI) + 90, in order to fit two curves between 2001 and 2009. Two major spikes in the COP curve are related to stock splits in 1985 and 2005.
Figure 4. Historical (close) price for COP (black line) and the difference between the core CPI and headline CPI between 1970 and 2009. The latter is scaled by the following time dependent function (+4)*(core CPI-CPI) + 0, between 1980 and 2001; (-4)*(core CPI- CPI) + 90, between 2001 and 2009.
Figure 5. Comparison of the observed and predicted price for COP stock between 2000 and 2009
There exist also "close price adjusted for dividents and splits". Let's consider it as a new time series to model. Preliminary results are presented in Figure 6 and 7 . The adjusted time series demonstrates much better fit between the observed and predicted prices. There is no changes to the text in the previous post. The only correction is related to the empirical relationships. Now:
COP =(-6)*(core CPI- CPI) + 80 ; from 1998 to 2009
COP = (+3)*(core CPI- CPI) - 5 ; from 1980 to 1998
COP = (+3)*(core CPI- CPI) - 5 ; from 1980 to 1998
So, the next trend guarantees the rate of COP growth two times larger than that of the difference between the core CPI and the headline CPI.
Figure 6. Historical (close) price for COP (black line) and the difference between the core CPI and headline CPI between 1980 and 2009.
Figure 7. Comparison of the observed and predicted price for COP stock between 2000 and 2009
Comment