Here we revisit our
price model for AutoNations’ (NYSE: AN), which we have been modeling since
2009. In March 2012, we presented a
brand new prediction for AN; we expected no large changes in the first
half of 2012. At least the price was not expected to go out of the uncertainty
bounds defined by monthly high/low prices. This prediction was right. The AN (monthly
closing) price was hovering around $35 before it started to grow in July and
reached the level of $44 in September 2012. Our current model shown the price
to fall back to $35 in the months to come. The closing price for the 12

^{th}of November was $40.56, i.e. approximately 10% down from the October’s closing price. The model suggests that no recommendation to buy AN should be given and there is no reason to keep it in the short run (few months). We will revisit the model in two-three months.
AutoNations’
is a company from services sector (as defined by the S&P 500) and operates
as an automotive retailer in the U.S. here we present the current model which
has been obtained by decomposition of the time series of monthly closing share
prices (adjusted for splits and dividends) into a weighted sum of two consumer
price indices. One might presume that a fast growth in the CPI inherently
linked to the AN share price (e.g. energy consumer price for energy companies)
relative to some independent by dynamic reference should be manifested in a
higher pricing power for the company. Therefore, the task is to find two best
(say, in sense of RMS residual error) defining CPIs. It allows testing of the underlying
concept (decomposition into CPIs) and to estimate time lags and coefficients
for AN.

We
have borrowed the time series of monthly closing prices of AN from Yahoo.com
and the relevant (seasonally not adjusted) CPI estimates through September 2012
are published by the BLS. The evolution
of AN share price is defined by the consumer price of rent of primary
residence (RPR) and the index of financial service (FS). In March 2012, the defining
time lags were as follows: the RPR index leads the price by 10 months and the FS
index leads by 5 months. The revised model has only one difference, the lag of
RPR is one month longer. The relevant best-fit models
for

*AN(t)*are as follows:*AN(t) =*-1.89RPR

*(t-10) – 0.36FS(t-5) +*15.47

*(t-*1990

*) +*270.31, February 2012

*AN(t) =*-1.95RPR

*(t-11) – 0.37FS(t-5) +*16.08

*(t-*2000

*) +*439.34, September 2012

where

*AN(t)*is the AN share price in U.S. dollars,*t*is calendar time. This model is valid since August 2011 with the same lags and coerffcients. Figure 1 displays the evolution of both defining indices since 2002. Due to the negative coefficient (slope), the sharp drop in the FS index in 2009 best explaines the jump in the share price five months later.
Figure 2 depicts the
high and low monthly prices for the share together with the predicted and
measured monthly closing prices. The predicted prices are well within the
bounds of the share price uncertainty before June 2012 and then the curves
deviate. This deviation has been growing and presented the challenge to the
model. It also supposed that the probability for the price to fall has been increasing
with time. The model residual error is shown in Figure 3 with the standard
deviation between July 2003 and September 2012 of $2.01.

Both CPIs have
negative influence on the share price. Therefore, the price should decrease
when the indices grow fast.

Figure 1. The
evolution of the index of rent of primary residency (RPR) and the index of financial
service (FS).

Figure 2. Observed and predicted AN share prices.

Figure 3. The model
residual error: stdev=$2.02.

## No comments:

## Post a Comment