Wednesday, December 23, 2015

Mattress Firm (MFRM) - Earnings Quality Shaky at Best


Mattress Firm (MFRM) is a mattress retail roll-up story that has doubled in value since coming public in 2011. MFRM is a posterchild for earnings quality analysis. A deep dive into the company’s accounts shows a reporting approach that is at best promotional, and at worst misleading.

 
A Serial Acquirer

In reporting its quarterly operating results, management makes liberal use of a wide variety of non-GAAP financial metrics, including (but not limited to) Adjusted EPS, Adjusted Cash EPS, EBITDA, and Adjusted EBITDA. The business model is fairly simple, so it’s not quite clear why so many varied metrics would be necessary in order to properly understand the business. Part of the rationale appears to be that the ‘Adjusted’ metrics leave off acquisition-related costs (as well as many other items). The only problem is that when you’ve done this many acquisitions (see below), then that is simply a part of your (possibly your entire?) business model. It’s hard to make the case that those costs should be excluded in order to properly reflect the business.

 

source: company presentation November 2015


Economies of Scale?

Presumably, MFRM’s roll-up strategy is predicated on the ability to extract value from the newly acquired stores over time. Possible sources of value creation could include cost reductions, revenue synergies, or opportunities for attractive CapEx investment. But a comprehensive review of MFRM’s operations calls into question whether the company’s acquisition roll-up strategy is actually creating value over time. A careful analysis shows that most of MFRM’s operational metrics have actually deteriorated since the 2011-2012 time frame. Whether via management discretion or purely incidental, MFRM’s business metrics built momentum and peaked shortly after the IPO in late 2011.

The first interesting metric is comparable store sales. This concept is an important one for all retailers. In some form or another, the purpose is to measure the performance for stores versus the year ago period for all stores that were open during both periods. It is designed to portray the underlying business performance exclusive of new stores (whether acquired or built). MFRM has a convoluted definition for its comparable store sales measure (see p. 35). Within its calculation of comp store sales, MFRM includes sales from relocated stores, as well as online sales. Management regularly reports and provides guidance on this basis. Yet the 10K also includes a measure called “average net sales per store unit”. Despite not being the metric that management regularly touts, it is actually closer to the expected definition of same store sales, capturing “sales for stores open at both the beginning and the end of the period, excluding e-commerce and multi-channel sales”. The chart below shows the trending of both of these metrics over time.

 
Notice in the final column that MFRM’s reported comparable store sales metric consistently exceeds the sales per store unit metric. Mechanically, this means that relocated stores and e-commerce sales serve to benefit the reported comps metric, although the company does not disclose any breakout of those components. Also note that the company has not reported a negative comp store sales year since the IPO in 2011. Yet the more conventionally calculated metric was negative in 2013.

Next we will go through the trends in various metrics for the business, using both the company’s preferred metrics as well as other conventional measures of profitability.

 

 
The first thing to note in the chart is the growth in store count. We know that MFRM is a retail roll-up story, but the sheer growth in stores is amazing. The company’s store footprint has grown over 5x since 2008. And the recently announced acquisition of Sleepy’s will add 1,066 stores. Despite this massive growth in scale, though, we are seeing deterioration in per-store economics and stagnation in company margins. The bottom of the chart shows the trend in management’s preferred metric of Adjusted EBITDA. 2014 marked the lowest Adjusted EBITDA margin since 2008, while Adjusted EBTIDA per average store had fallen back toward 2010 levels. This trend flies against the notion that MFRM’s massive scale expansion is leading to economic benefits.

If we then switch over from the company’s preferred metrics to more conventional measures, note that gross margin has basically stagnated for four years, after reaching 39%+ levels in 2011. Similarly, SG&A expense leverage has actually gone the wrong direction. The exponential rise in store count has resulted in SG&A % of sales approaching 33%, up from 29% earlier in MFRM’s history. Again, this is a discouraging result for believers in the company’s scaling roll-up strategy. On a per-store basis, MFRM earned only $57k per store in FY 2014, net of COGS and SG&A – a level not seen since 2009.

The next chart compares these profitability measures to MFRM’s cash flow trends.

 

Similar to the SG&A margin trends just mentioned, the chart shows that MFRM’s operating cash flow per average store in FY 2014 fell 30% to levels not seen since before 2010 – only $63k per store. Again, whether we use true operating cash flow or the company’s preferred Adjusted EBITDA metric, the trend is basically the same. As a final assessment of MFRM’s operating efficiency gains (or lack thereof) due to scale, the next chart shows the trend in cash conversion cycle.

 
 

In addition to margin stagnation and deterioration in per-store economics, MFRM’s cash conversion cycle trend is an indictment of the company’s massive expansion strategy. In 2008, the company was receiving cash from customers 9 days prior to paying its own suppliers. Fast forward to 2014, and we see a complete reversal of nearly 4 weeks, as payment from customers comes 18 days after the sale. That might not seem like much, but the directional trend is very concerning.


Valuation

All of the above is easily enough for me to consider MFRM as a short-sell investment, given the right price. So what is the right price? Well we know that MFRM just announced their pending acquisition of Sleepy’s. Their presentation shows the following valuation statistics for the transaction.

 
source: company presentation November 2015

So MFRM is paying $780 million for 1,066 stores – equating to an enterprise value per store of $732k. Since MFRM has a strategic interest in acquiring Sleepy’s (given synergy potential), it is logical that MFRM would value Sleepy’s higher than any other potential buyer. This provides a fresh, relevant data point for current market valuations.

The chart below shows historical valuation levels for MFRM on a per-store basis. The 2016 estimated store count gives effect to the 1,066 additional Sleepy’s stores (and the additional debt), and backs into the implied stock value if MFRM as a whole were valued near the $732k enterprise value per store.
 


So why would market enterprise values drop to the levels indicated in the chart? Well the process has already begun, and I believe the answer is leverage. MFRM shares are poised to end 2015 around $45. This would mark the first calendar year end for MFRM’s valuation to fall below $ 1 million per store. Given the increasing debt load and general credit conditions, the Sleepy’s transaction valuation level around $730k-$735k EV per store does not seem unreasonable. This would imply a $30 stock price, or roughly 33% decline from current valuation levels around $45.

 
Disclosure: author is currently short shares of MFRM

Monday, December 7, 2015

First Thing's First

Earnings Quality – what is it and why is it important?

          I’ll take it in reverse: EQ is important because earnings are the basis for valuing a stock. ‘Earnings’ here does not refer to net income specifically, but rather to whatever measure is used for valuing annual economic performance. Net income is a pretty bad measure for this, because it is impacted by and susceptible to such a wide variety of factors that are not going to be indicative of the underlying, ongoing economics of the business. Better measures include free cash flow to the firm (FCFF), net operating profit after-tax (NOPAT), or even operating cash flow less CapEx (if you want to rely on a third party source and not do any calculations yourself).
          So what is earnings quality? Quality of earnings measures the degree to which reported operating results accurately reflect the economic performance of the business. Put another way: to what extent can we accept management’s reported results at face value? There are many ways to assess and measure earnings quality (and those will be discussed extensively here in the future).


Application – micro and macro
          From an individual company perspective, the application for earnings quality analysis is fairly obvious. The process is basically:

          - how did reported results compare to consensus expectations?
          - were there earnings quality issues that could account for the positive or negative surprise versus expectations?
          - does the combination of stock price movement due to earnings surprise and the presence of earnings quality issues provide an opportunity to make money (long or short)?
 
          Another more subtle implication is that management intent is basically irrelevant for this analysis. For example, whether the quarterly reported sales exceeded consensus estimates due to aggressive accounting assumptions by management or due to the (appropriate) application of a new revenue recognition standard that was not fully understood by consensus views still drives the same conclusion for earnings quality analysis. Namely, the recently reported results were aided by an accounting or reporting phenomenon that was not discounted in the consensus view. This then requires assessment of whether such phenomenon is recurring and sustainable, or whether it constituted a non-recurring effect on results (which by definition is likely to reverse subsequently).

          From a macro perspective, I think there is an important and under-appreciated long-term trend affecting comparability of stock indices (such as the S&P 500) over time and the respective P/E multiples applied to them. The proliferation of non-GAAP earnings measures being reported by the vast majority of public companies currently creates a comparability issue with historical valuation measures. Although I have not done the formal work (not enough hours in my day), the hypothesis is that a larger and larger proportion of companies now utilize and report (and perhaps more importantly, Wall St sell-side analysts now base their coverage and estimates upon) some variation of ‘adjusted’ or ‘non-GAAP’ earnings. Increasingly, this typically just represents ‘earnings without the bad stuff’. Even when adjusted EPS appropriately only removes truly non-recurring items, the issue of comparability with historical periods still persists. This makes comparison of today’s market P/E multiple to that of previous cycles less and less relevant.


It’s a market of stocks

          So why do earnings quality analysis? Because the stock market is not (necessarily) the index. Sure, the right answer for many (most?) individual investors is to index. But for professional investors or individuals with the skill set, it is important and potentially very profitable to understand that it is a market of stocks – each an individual company with specific factors affecting value.

          By definition, any index is going to include a lot of losing stocks. Analyzing earnings quality is an important tool in identifying and avoiding (or shorting) those. The actual numbers might surprise you. Here is but one of many examples showing how high a proportion of stocks in the index are actually long-term losers, and by definition, how the index’s positive long-term returns end up being due to a select few stocks that massively outperform. This particular study looked at the period from 1983-2007, examining 8,054 stocks that would have qualified for membership in the Russell 3000 at some point in their lifetime. 39% of these stocks produced a negative lifetime total return, and 18.5% declined by 75% or more. 64% of stocks underperformed the Russell 3000 index. The bottom 75% of these stocks collectively had a total return of zero, while ALL the gains were attributable to the top 25% of stocks. “In other words, a minority of stocks are responsible for the majority of the market’s gains.” Del Vecchio mentions this study in his book, which is highly recommended.