Comp Analysis - Questions ?

I was wondering if some one could guide me on how comps are built at a BB.

  1. Is the data pulled from financial reports or is it taken directly from CapitalIQ, Factset, etc ?
  2. What all has to be done to scrub the comps ? Are steps mentioned below all that is needed ?
  • Normalizing Earnings (Adding one off / restructuring expenses and subtracting one off or non recurring gains )
  • Adjusting for different financial periods(31st March vs 31st December)
    -Converting information to LTM
    -Adjusting Enterprise Value (EV) by removing the expense or income related to under or over funded pensions
    -Value Employee Stock Options & Preferred Stock
  • Adjust EBITDA for income from Strategic Vs Financial Investments
  1. Are the froward EPS Estimates (2008E, 2009E, etc) an average of those obtained from few brokerage reports, or concensus analyst estimates from reuters or bloomberg ?
  2. Are there any other multiples included in the comps other than : P/E, EV/Sales, EV/EBITDA, EV/EBIT Price/BV, ROE,5 year CAGR, PEG, etc
  3. Do analysts in Industry groups usually have a set of comps ready to be used in a pitch ? Otherwise..How long do comps take to build on average ?
  4. What about precedent transactions ? Are they taken directly from factset, dealogic, bloomberg or CapitalIQ ?

Comments (46)

Jun 30, 2007
netking007:
  • Adjust EBITDA for income from Strategic Vs Financial Investments
  1. Are the froward EPS Estimates (2008E, 2009E, etc) an average of those obtained from few brokerage reports, or concensus analyst estimates from reuters or bloomberg ?
  2. Are there any other multiples included in the comps other than : P/E, EV/Sales, EV/EBITDA, EV/EBIT Price/BV, ROE,5 year CAGR, PEG, etc
  • For earnings estimates it's pretty standard to use the I/B/E/S consensus estimate on the bloomberg EEO screen. I've never seen it done otherwise.
  • Depends on what industry you're working with. There are many specialized ratios like EV/reserves or EV/mboed for oil and gas producers, or TAC/users (traffic acquisition cost) for internet portals.
  • I don't get your strategic v. financial EBITDA Q. \Instead of starting with NI and adding back ITDA, use operating income (and make sure they don't have D&A above that line) which ignores earnings from JV's and the like.

    Jun 30, 2007

    1:
    From what I've seen from interning (boutique), the answer is to get the number right. Hence, I usually go the route of direct financial statements, particularly because CapIQ is wrong so frequently (especially with smaller companies).
    2:
    You have to do more than just figure out the LTM of data, you also have to properly calenderize it. Comparable companies can have different fiscal years, throwing off your future estimates from analysts who state estimates in the respective company's FY.
    6: See #1. You have to get the #s right. What I do is pull the data from a source, then compare it with press releases obtained from company websites. If there is a discrepancy, I stick with the website #s. This way I am double checking my work against itself and against multiple sources.

    Jun 30, 2007

    1) Always from 10Q/K, unless specifically told otherwise (quick and dirty calc)
    2) Need to really only adjust for historical comps. Adjust for non-recurring, etc.
    3) I/B/E/S consensus mostly, sometimes from ER reports
    4) The multiples you use depends on your industry (telecom uses ARPU, etc), but the ones you mentioned are most common
    5) There is usually a comp set, but its a different story whether its kept updated or not. Comp time depends on how many companies, etc. Historical comps can take a long time with all the adjustments.
    6) Precedents are a bitch- get info from anywhere you can- Factset, PR, website, etc.

    Jul 2, 2007

    Thanks for all the information. Any thoughts on the below questions :

    Q1. How do stock options & preferred stocks get factored into a comp - analysis ?

    Q2. For historical comps...I know that one time expenses (from restructuring) are added back while one off / non recurring gains are subtracted. Hence changing the EBITDA & EBIT numbers.
    The Net Income is also adjusted taking marginal tax into account. Hence i am assuming the EPS number changes as well...making these adjustments ?

    Q3. While creating a comp...the ratios for how many historical & future years are needed ? For example for a comps being created in 2007....I would assume the historical years would be 2004 - 2006 and the future ratios would be for 2007 - 2009 ?

    Thanks in advance.

    Jul 2, 2007

    Correct answers to your questions would depend on the purpose of your analysis.

    1. Value and deduct the stock options from equity value for regular stock analysis. Also take into account that not all options would be exercised, so it's important that you have a historical ratio of exercised options. If this is for a live deal all outstanding options would be immediately exercised due to change of control provisions that call for immediate vesting of all options and restricted stock.... so the buyer would foot the bill.

    2.Don't understand your question, but I'd simplify things and only look at operating income, excluding all one time gains and charges.

    1. P/E multiple is based on stock price and ttm EPS unless otherwise stated. The trailing twelve month earnings is all you need for "history".

    If you are valuing a cyclical firm, normalize earnings and other inputs over the last cycle. That way you don't under or over value your target in a good or bad year.

    Learn More

    7,548 questions across 469 investment banks. The WSO Investment Banking Interview Prep Course has everything you'll ever need to start your career on Wall Street. Technical, Behavioral and Networking Courses + 2 Bonus Modules. Learn more.

    Jul 3, 2007

    Following from the above comments..I have a quick question :

    If P&L statement has the following line items:

    Sales
    -Cost of Goods
    Gross Profit
    -Operating Expenses
    EBIT
    +D&A
    EBITDA
    + Total Non-operating Income
    - Total Non-operating Expenses
    Pretax Income
    - Income Tax
    Net Income
    - Dividends
    Retained Earnings

    To calculate EPS...I am assuming the Non-operating Income needs to be subtracted from the Net Income and the Non-operating expense needs to be added to the Net Income. Then the Net Income is divided by the total number of shares to get the corrected EPS.

    Is the above correct ?

    Jul 3, 2007

    Unless otherwise stated, reported EPS does not care about the source of income, so whether it's from direct operations, or dividends from a subsidiary is really irrelevant. If you look at a 10-Q carefully, you'll see only one line for Net Income Applicable To Common Shares which is after all sources of income and their expenses (cash & non-cash).

    During a restructuring the firm can also report "EPS excluding restructuring charges", so investors see the net effect of restructurings on EPS.

    Jul 3, 2007

    So....in that case the only time you need to make changes to the Net Income is when the income is a one-off (non-recurring)..and the nature of the income(core / non-core) would result in no changes to the Net Income and hence EPS.

    Is that correct ?

    Jul 3, 2007

    Personally, I would take a step back and intuitively understand the business profile of the company you are looking to value. That requires you to go beyond an industry cut, and look at things like sub-sector, customers/end markets, distribution channels, and products and services.

    Then I would look into the financial profile of the company and make cuts based on the most relevant metrics for that type of company. If looking at a oil & gas company, production levels are more important than revenues and looking at industry specific metrics like EBITDAX is more relevant than EBITDA.

    But never overlook asking an experienced person in your team if they have a list of comps on hand already (you may need to polish it a bit of the list is over a year old). That is usually the "best" and most efficient approach.

    Jul 3, 2007

    I'm looking at a tech company so I assume revenues or EBIT are most relevant. The company is early stage and doesn't have a high EBIT so I used revenues. Still don't know how to further narrow down.

    Also, can someone please provide insight in regards to the second question?

    Jul 3, 2007

    I don't think you should think about this in terms of absolute values (ie: finding companies with $20M revenues versus $75M). At that level, it is hard to find good public comps.

    Clearly, you're looking for high growth tech companies in your sub sector. I would screen for something based on your growth metric, so 75-100% YoY Revenue Growth. You can tailor this screen depending on how many hits you get and how meaningful they are. Remember, companies with higher growth prospects trade at higher multiples.

    Since your company is in its growth stage, odds are it is unprofitable, or has very low earnings. As a result, the only multiple you can really use based on current and next year's numbers is a revenue multiple. However, you should still spread EBITDA, PE, and PEG multiples for your comps, since you are likely projecting the company's financials 5-10 years going forward, which will show predictable earnings at some point, thus making the profitability metrics relevant for any exit / M&A scenarios.

    Based on where the comp group trades on an NTM and LTM basis, you should be able to come to a good range.

    Jul 3, 2007

    Thanks so much for the help guys. I thought I understood a lot of this conceptually, but actually doing it I've realized how little I know.

    A somewhat stupid question, but after I spread the multiples and decide to use say a revenue multiple - do I multiply that by 5th year EV to get to terminal value?

    Jul 3, 2007
    arguewithatree:

    Thanks so much for the help guys. I thought I understood a lot of this conceptually, but actually doing it I've realized how little I know.

    A somewhat stupid question, but after I spread the multiples and decide to use say a revenue multiple - do I multiply that by 5th year EV to get to terminal value?

    you are getting a lot of things mixed up.

    when spreading comps you are trying to ascertain an implied enterprise and/or equity value by using trading multiples from the comp universe you determined. so you use the multiple (for example, the median industry EBITDA multiple) and apply it to the relevant metric (in this case EBITDA) of the company you are analyzing to derive an implied enterprise value. but multiplying EV to get a terminal value doesn't make any sense, since a terminal value, technically speaking is the company's enterprise value in perpetuity.

    to get to a company's terminal value (usually for a DCF, not really for comps per se), you multiply your last projected year financial metric (lets say EBITDA in yr 5) by the terminal year EBITDA multiple. once you discount the product at the appropriate discount rate, then you've technically arrived to your terminal value.

    hope that makes sense.

    Jul 3, 2007

    Okay so I guess I'm calling it the wrong thing. I'm referring to terminal value in a DCF. To get the terminal year EBITDA multiple, can't you use comparable companies? Is that called something else?

    Jul 3, 2007

    EV is a $ amount... but generally speaking, you're assumption is correct.

    Jul 3, 2007

    Correct....

    Jul 3, 2007

    I would use the median and eliminate any crazy outliers, but yeah you've got the idea.

    Jul 3, 2007

    Thanks for the clarification.

    Jul 3, 2007

    No, GAAP EPS will include non-recurring items. To apply a consistent metric to be used to compare companies across the industry, an adjusted EPS to exclude non-recurring is a preferable measure.

    Jul 3, 2007

    You have to make necessary adjustments as stated above.

    Dirk Dirkenson:

    Shut up already. Your mindless, reflexive responses to any critical thought on this are tedious. You're also probably a woman, given the name and "xoxo" signoff, so maybe the lack of judgment is to be expected.

    Jul 3, 2007

    Target...

    If you're looking at transaction multiples, you're concerned about the price being paid for the target, not the trading valuation of the acquirer at the time...

    Jul 3, 2007

    Nice, nice....

    Do you guys not have the ability to look at even the most simple book / website that would tell you this? Why the hell would you look at the acquirer's metrics when doing transaction comps? This is common sense...

    Jul 3, 2007

    You could search the Amadeus database, but you probably dont have access to this so im outta ideas.

    Big 4 Accounting Guide to Getting Hired Contains interview questions, exactly how to answer, resume guide, how to make an impact and a guide to the firms and service lines.

    Jul 3, 2007

    is this a credit card processing firm?

    Jul 3, 2007

    sounds like this company does transaction processing. look into trade journals. also, it's possible to look up company profiles on linkedin. It's sometimes really helpful. tell me how it goes.

    Jul 3, 2007

    Payment processing:

    GPN
    TSS
    CYBS
    EEFT
    HPY

    Core processing:
    FISV
    FIS
    MV
    JKHY
    ACIW
    FNDT

    Jul 3, 2007

    just tell them you used capiq, but you understand the thought process behind comps and can answer theory questions.

    no one really hand spreads comps unless you're on a live deal anyway.

    Jul 3, 2007
    kidflash:

    just tell them you used capiq, but you understand the thought process behind comps and can answer theory questions.

    no one really hand spreads comps unless you're on a live deal anyway.

    Know any example theory questions I should know? Thanks!

    Jul 3, 2007

    My VP makes us hand spread all comps on a quarterly basis for the 35-40 public companies in the industry we follow. fun stuff

    Jul 3, 2007

    I would study 1) methodology you would use to screen and refine your list of comps comps, 2) what metrics to spread and why they are important, 3) unlever and relevering the beta, 4) where comps falls generally in the valuation spectrum (i.e. does it throw off a higher valuation than a DCF, LBO, precedent transaction analysis).

    All of the above can be googled quite easily.

    Jul 3, 2007

    I think you get S/O from the 10k and then just use TSM to get diluted # of shares

    BossMode

    Jul 3, 2007
    buttercup87:

    I think you get S/O from the 10k and then just use TSM to get diluted # of shares

    But what about NCIBs, equity offerings, options granted, etc..? It would seem kind of odd to use the 10K and then make adjustments for these situations.

    Jul 3, 2007

    but i think u can get the number of s/o from either k or q, then you can find the other in proxy too...

    Jul 3, 2007

    Can someone please give me a definitive answer on how to do it properly?
    I can't seem to find the answer anywhere.
    Thanks.

    Jul 3, 2007
    Comment
    Jul 3, 2007
    Jul 3, 2007
    Comment