Benchmark flaws?
In light of the continued robust performance of the Russell 2000 Value Index™, and the other Russell Value Indexes, we hope this letter helps our clients and investors better understand that there may be imperfections in the indexes that may distort return comparisons. We will discuss the methodology Russell uses to construct its value indexes, give examples of some companies that conflict with how we define value, and hope to bolster our contention that these value indexes should not be used as benchmarks for absolute value managers, such as ourselves.
The Russell Investment Group uses a probability methodology to allocate stocks to either the growth or value indexes. Basically, this method allows for companies to have both growth and value characteristics, based on a relative book value-to-price and long-term earnings forecasts. Essentially, this method forces companies in the Russell indexes into either the value or growth index or, counter-intuitively, both indexes. Additionally, Russell assigns a weight to each security within the relevant index that is intended to represent the market and manager behavior. Nevertheless, as we’ll discuss below, Russell’s style methodology uses a relative value approach that is not in harmony with our absolute value style.
Alas, we believe the year-to-date returns of the Russell 2000 Value Index, and the other Russell Value Indexes, have been skewed by a number of companies that do not warrant the “value” label. A reason, among several, is that the various Russell Value Indexes include companies that have not earned any profits in all the years that they have been publicly traded.
One of the more striking examples of a company that is used to help construct the current Russell 2000 and 3000 Value Indexes is Nuvelo, Inc. (NUVO). Nuvelo, which appreciated 119% through the first quarter of 2006, is a biopharmaceutical company focused on the discovery, development and commercialization of novel products for acute cardiovascular indications and cancer.
NUVO has been a publicly traded
company since 1997 and over that time frame has “achieved” cumulative sales of
just $91 million and over $300 million of losses. Within
the last twelve months, sales total a miniscule $545 thousand, while losses are
$71 million.
Surprisingly, Wall Street’s
consensus earnings forecast for NUVO is continued losses for this year as well
as through 2008.
As for its book value, NOVU is
burning cash faster than a wildfire on a hot, summer day in
Despite the fact that NUVO is included in the Russell 2000 and 3000 Value Indexes, this company does not meet any of our criteria for a value stock. We would describe NUVO as a publicly traded, venture-capital stock. In our opinion, NUVO should be characterized as a speculative stock without any attributes that convey value.
So, how does a company like NUVO find its way into the Russell Value Indexes? In NUVO’s case, the company’s share price declined more than 40% just prior to Russell reconstituting its indexes in the middle of 2005. The share price was falling because, among other reasons, investors were concerned about how much money NUVO was losing.
In order to sustain its operations, NUVO sold nearly 10 million shares of stock in the first half of 2005, for approximately $72 million, which more than doubled shareholder’s equity, or its book value. The combination of a low share price and larger book value allowed NUVO to meet the Russell 2000 Value’s book value-to-price metric, despite the company’s enormous losses and paltry sales.
At FPA, we use a company’s historical sales, normalized earnings, book value, and cash flow to determine whether a stock represents good value for our client’s capital. If a company has not earned any profits for its shareholders, nor has any tangible assets that are valued below current market prices, we would not consider such companies as value opportunities. In other words, we use a fixed, low multiple of sales, cash flow, and earnings to determine value. We do not endeavor to subject our clients and shareholders to the risk of playing the relative value approach, which, in the extreme, could lead to absurdly valued companies.
Thus, we have a philosophical difference of opinion about the parameters used by Russell to construct its value indexes, as an appropriate benchmark for a value manager like FPA. Like any complex issue, the details do matter. The basic issue is one of investing versus speculating.
In our opinion, the Russell Value Indexes include companies that have highly speculative characteristics. In addition to NUVO, the Russell 2000 and 3000 Value Indexes include a company called Atherogenics, which has not generated any revenues over the last four years and lost $80 million in 2005. The Russell 3000 Value includes additional companies that have similar results as Atherogenics Inc. – namely, Antigenics Inc. and Threshold Pharmaceuticals. We can also highlight companies in the Russell 1000 Value Index, for example, American Tower and JDS Uniphase, which have not achieved any positive earnings since being publicly traded or over the last decade.
Currently, over 6%, 11%, and 10% of the Russell 1000, 2000, and 3000 Value Index, respectively, include companies that have generated a negative 5-year return on assets or equity and are trading above 1x sales. In our opinion, valuations start to get rich when a company’s market capitalization rises above 1x sales. In the Russell 2000 Value’s case, the average first-quarter return in 2006 of these dubious, value companies was 23%, which has positively benefited the Russell 2000 Value’s return in Q1 2006.
It is not our intent to have the Russell Investment Group change their index construction methodology, but merely to highlight some inconsistencies that we believe the methodology may have with the investment style it claims to track. While the Russell 2000 Value was up an impressive 13.5% in the first quarter of 2006, we question whether that return accurately portrays a value manager’s opportunity set when the index includes “value” companies such as Nuvelo, Atherogenics, and many others. Thus, we believe it is wise for our clients and shareholders to understand that the Russell Value Indexes, as they are currently constructed, are not appropriate benchmarks with which to compare FPA’s performance.