Anyone want to learn intrinsic valuation? sure Yeah. As I've said, it would be nice to learn the correct 'book' way to do things rather than my (bad) experience based way. :) ok it will take a bit of time and energy are you all busy working? The only thing on my plate for the next 2 hours is getting the kids breakfast what I'd suggest we do is pick a company and value it Would EMC be a good test case (might reassure jlintz )? and in the process I will be able to cover most of the critical things I think as a starter we would be better off with a simpler company :) OK. we want to avoid things like cross holdings, lots of different business segments, etc. it would be good to pick a company that has a fairly straightforward business model SYY? (The other Sysco) Or you pick. BZH - home builder SYY works for me, though maybe you can take a moment to check if it has any cross holdings it looks like not, at a glance Yeah, I didn't see any. ok cool we will use a helper spreadsheet, which you can find at http://www.stern.nyu.edu/~adamodar/pc/fcffginzu.xls (assuming people have Exhell) Open office should open it too But I do have excel actually we should probably start with my helper sheet... Ok. (Reading the spreadsheet now) one moment, I'll try to put this somewhere where you can get it http://192.94.73.4:7740/FCFF-Data-Template-2.xls gents good day locklace: Got it Morning Yggdrasil ok, this is the sheet I actually use to collect the data the Primary tab contains all the info needed by the fcffginzu spreadsheet, in a form that makes it fairly easy to paste in it also has a couple of other items Cool, be back in <5 minutes. Kids are ready for breakfast. Go ahead, I should be able to catch up the Secondary tab is used to help compute some figures in the Primary tab when we want to use trailing twelve-month (TTM) numbers, i.e. the latest ones available. (which we always do) if the time of year is such that the most recent quarterly filing is actually a 10-K, then we can avoid filling out most of the Secondary sheet but in all other cases we need to fill it out completely so we have the TTM figures for the firm. Ok. (Back.) we can get most of what we need from one 10-Q and one 10-K (the company's latest 10-Q and latest 10-K) when did SYY last report earnings? let's see Their last 10-q was on Nov 8 ok, they have a weird FY schedule Yeah. so this is, what their First Quarter Q? For Jul-Sep Yeah. yes, but it doesn't seem to have 3-month numbers! this is what I mean about complete lack of standardisation in these filings oh, never mind I was reading 13-week as 13-month...haha Heh. I was looking at it thinking that I wasn't sure what you weren't finding. (everyone else on the planet says "3 months", and these guys say "13 weeks") Their CFO might be female and thinking of it the same way as she would a pregnancy. :) haha "William J. DeLaney III" hope not :) ok, anyway, back to the sheet So we can put in "2007-09-30" and "2006-09-30" under This Period and Next Period on the primary sheet, so we know what timeframes we're working with BTW, It does look like they did a share repurchase during the quarter. now let's talk for a moment about R&D and operating leases. that's fine, that will come out in their book value of equity When we work with financial statements, we have to 'normalise' them into finance terms because they are produced by accountants and accountants are often very silly. heh One thing that accountants do that is very silly is require that R&D is accounted for as an operating expense, rather than as a capital expenditure. If you think about it for two seconds, you will agree that R&D, far from being an opex, is the ultimate capital expenditure. I imagine they do that since R&D is at least somewhat tax advantaged. or the other way around - opex is a tax benefit, so we'll put R&D there. ;) Problem is, from a valuation perspective it screws things up because R&D really is a capital expense, and it's critical that we have as good a sense as possible about how much the firm is reinvesting. So what we'll be doing in valuation is reclassifying R&D expenses as capital expenditures (and taking care of all of the normalisation that requires; the spreadsheets will do this for us automagically). How much difference this makes obviously depends on how much R&D the company is doing. Not much in this case Another thing we will do is classify operating leases as debt, because in effect that's exactly what they are. This does not usually make much difference at all, except for certain businesses - in particular for brick-and-mortar retailers like The Gap it makes a huge difference, since most of their "debt' actually takes the form of operating leases (for stores). Now I need to back up again and talk about what we're doing. We will be doing something called "intrinsic valuation", otherwise known as "discounted cash flow (DCF) valuation." There are two basic kinds of intrinsic valuation. We can value the entire business, or we can value just the equity in the business. What you did with WFR. The first approach is called the FCFF approach - free cash-flow to the firm. The second is called FCFE - free cash-flow to equity. In a sense from an equity-investment point of view it doesn't really matter which approach you use; it's more a stylistic choice than anything. The two approaches in principle should yield the same result for the value of equity in the business. There are some general advantages to using FCFF in many cases, though, and we will use FCFF here. (I rarely find myself using FCFE.) ok So in FCFF we will be coming up with a value for the entire business, which includes both the value of debt and the value of equity in the business. In FCFF we tend to think of creditors as being partners with equity investors, a viewpoint that as equity investors we don't often take. :) Certainly not when a company goes bankrupt Both provide capital to the business and expect a certain return, so really they are both partners in the business, although the nature of the risks they each assume is quite different. Indeed. Now, intrinsic valuation is different from the other primary approach used for valuation out there in the world - *relative* valuation. In relative valuation we are asking a different question: not "is this company overvalued or undervalued with respect to its intrinsic value" but "is this company overvalued or undervalued with respect to *the way the market is currently valuing similar firms*". ice hole swimming in one hour, woot mortal: don't forget to leave us the contact info for your next-of-kin Crazy Finns. It's entirely possible for a company to be overvalued on an intrinsic basis but undervalued in relative terms, for instance. This often happens when the market is overvaluing an entire group or industry. Relative valuation usually takes the form of "valuation by multiple" - price/earnings, price/sales, price/book, Enterprise value/EBITDA, EV/Sales, etc. Relative valuation is problematic, not out of necessity but because it's usually done very badly. ok i.e. without a grasp of even the most basic statistical rudiments (like when it makes sense to ask about an "average value") and without an understanding of intrinsic valuation. We take the view that intrinsic valuation is the backbone of all valuation. If you understand IV, you can easily do (correct) relative valuation, but the converse is not true. Intrinsic valuation is to finance what anatomy is to a surgeon. got it ok, so with that in mind, let's start filling in the grey boxes in the primary sheet. first let's do R&D - do these guys even do R&D? I didn't see any. The Summary 10-Q on yahoo had a "-" line item. From the 10-K: "We are not engaged in material research and development activities relating to the development of new products or the improvement of existing products." Guess that's that. Not unexpected. :) You did want me to find a simple example. ;) So go to the R&D tab, and fill in 0 for the first six grey boxes and back on the primary sheet, put a 0 in for Current Year ok Now for operating leases. This info is on p.45 of the 10-K. oh, one other important point: we will input *all* values in millions! so "Current Year" for Operating Leases should be 94.163 Was wondering about that. And what site do you use to grab the 10-k/q's? Seems the Yahoo/Edgar partnership isn't properly downloading the whole thing for me right now. oh, use sec.gov http://www.sec.gov/edgar/searchedgar/companysearch.html Ok, loaded. So our operaring lease data should look like: 94.16, 63.38, 53.32, 45.24, 36.20, 27.27, and 142.30. ok Now: what should we use for the tax rate? There are basically two choices: we can use the effective tax rate, or the marginal tax rate. What do you think? I'd lean towards effective tax rate I'd imagine it to be less work than calculating the marginal rates. ok, we actually want to use the marginal tax rate for valuation in general. Why? The reason is that the effective tax rate tends to be very volatile. It goes up and down a lot as a company manipulates their various tax-valuation allowances and deferral credits from year to year. So the effective tax rate you see at any given time is not necessarily a good indication of how much tax the company will be paying over the long term. Makes sense. The marginal tax rate is a country thing, so you can just look it up (although it can be hard to find). e.g. in the US it's 35% and maybe another 5% for state tax. * joshin grumbles about how that high of a rate hurts US Business Now there are actually two places where the tax rate will show up in our valuation. One in the high-growth phase and the other in the stable growth phase when calculating terminal value. We can use different rates in these two phases, and it often makes sense to do so. ok If you go to Reuters and look up a stock, and then hit the Ratios page, you'll see figures for current and 5-yr avg. effective tax rate in there. 39.18/38.24 /s/9/8 There are a couple of ways to do this, but what I'd suggest is using the 5-year average effective tax rate for this bit, especially since it's higher than the marginal rate. So we'll put in the 38.24% here. ok Next is working capital. Here again we'll diverge from the accountants. What's the accounting definition of working capital? Isn't it day-to-day capital availability? * joshin goes to wikipedia If I wasn't talking in accountant speak, I'd think it was the amount of capital the company is using to generate revenue. Aah, that's "operating capital" Accounting definition of WC is 'current assets - current liabilities'. WC is meant to represent, in essence, how much a firm is reinvesting in the short term, i.e. for the current period. Ok. When we talk more about reinvestment, we will see that it amounts to net capex (capex - depreciation) + change-in-WC. which is in essence long-term + short-term reinvestment. ok but in order for this to make sense we need to normalise WC. What we will actually use as our definition of WC is "non-cash current assets - non-debt current liabilities". The reasons for this are a bit subtle, but essentially it's because cash is not a wasting asset - it's going to be, in practice, earning a fair return, since the "cash" you see on a company's balance sheet is not just currency stuffed in a drawer, but actually... what? Invested or in an interest bearing account Yeah, specifically it's going to be invested, usually, in something very safe like T-bills or commercial paper. and it's going to be earning roughly the risk-free rate. So for that reason we can net cash out of working capital. And as for debt, this will appear as part of the firm's book/market value of debt, so we don't want to record things like notes-payable/short-term debt and current-portion-of-long-term-debt as "reinvestment". so if you do this for SYY, you should end up with a WC value of 1471.20 for the current period. * joshin is trying to make the numbers match yours and 1425.94 for the prior period. a couple of mechanical notes on computing WC: we're computing (current assets - cash) - (current liabilities - debt). now "cash" will often appear as several line items on the balance sheet, e.g. cash, short term investments. sometimes there are two or three lines that have to be added up to get "cash". in the case of SYY, they just have one line item called Cash. Ok, numbers match now. for debt, there are usually two potential line items within current liabilities that apply: notes-payable/short-term-debt, and current-portion-of-long-term-debt. so you add up whichever of these appear to get "debt". and of course "current assets" and "current liabilities" always appear as their own line items so that part's easy. so as you can see the change in WC between these two periods is pretty low, only $45m. yup oh, one other mechanical note: most companies are not nice enough to provide the 'prior period' column on the 10-Q balance sheet. * evn_ (n=evn@pdpc/supporter/professional/evn) has joined ##finance instead you will see two columns, one for the current period ended, and the other as of the end of the last FY. so usually you need to pull the 10-Q for the same period a year ago to get the balance sheet. So I'd need to go back to the prior 10-Q and calculate that one. Oh right, prior year. yep. usually we will need that prior-period 10-Q for something else later on anyway ok, now you should be able to fill in the figures for Book Value of Debt and Book Value of Equity for the current and prior periods remember debt is long-term debt + short-term debt + current portion of long-term debt. So you should get: BV of Debt: 1976.08 and 1851.79, and BV of Equity: 3297.46 and 3187.84. So I don't need to know the book value of the company yet? * unix_infidel (n=blue@unaffiliated/unixinfidel/x-8383745) has joined ##finance Ok, apparently not. ;) ? I figured my own question out. Never mind. ok did you end up with the same numbers? Yeah. ok. next is cash & marketable securities, which we take straight from the balance sheet: 190.15 ok Value of Non-Operating Assets is the value of any holdings the company may have that aren't related to its operations. Often you'll just put 0 here, but if let's say you knew your company had $250m of luxury housing and art collectibles for senior execs, you'd put that in here Aah, the Tyco line item. in this case we'll just use 0 yep for Current Stock Price, let's agree on $28.31 and Shares Outstanding is 608.16 Where did you get that? shares? There's a line item of 610.8 and 617.11 (diluted) on the 10-Q I usually get that one from google Yup ok right getting the live figures should be more current ok okay, next is a weird-sounding one! "Unlevered Industry Beta" What could that mean and why would we care about it? It would show how SYY is expected to perform with respect to its peers when compared to the 5 year growth. A relativistic comparison. The "UIB" would show sector growth compared to the market. (At least in my little brain) Aha, found a definition. ok, it actually has nothing to do with growth Yeah, so I noticed. beta is basically our proxy for risk. lol, we're getting a lesson on valuation models and I missed it ;-P. Make sure to post a log or something. This gets a little bit complicated, but here's a quick summary of what the deal is with beta. One way to look at beta is as the degree to which a stock moves relative to a benchmark like the S&P 500. That's how I know it. A stock with a beta of 1.0 moves 1% for every 1% move in the S&P. A beta of 2.0 means it moves 2% for every 1% move in the S&P. A beta of -1.0 means a stock moves 1% in the opposite direction for every 1% move in the S&P. This kind of beta is called a regression beta and is computed by running a regression of a stock price against the market price over some period of historical time. For this reason you will hear people mistakenly refer to "beta" as "volatility". You will also see, for every financial page or publication on or off the net, a different "beta" value listed for the same stock. But "beta" is actually a parameter in the Capital Asset Pricing Model (CAPM), which we will not go into, but which gives us a way to define the cost of equity for a company as: r + b(e), where r is the risk-free rate, b is beta, and e is a number called the "equity risk premium" which we will talk more about later, but in essence represents the extra return demanded by investors when investing in the equity market, as a spread ... ... (premium) over the risk-free rate. ok so beta is a multiplicative factor - a higher beta is associated with higher risk and means a higher cost of equity. the question is, how do we come up with a value of beta for a company? and the wrong answer is, "use a regression beta." almost everyone in the investment banking world uses the wrong answer. Book value growth compared to risk-free rate? the main reason why a regression beta is an outstandingly bad way to get a beta is the standard error associated with the regression. the standard error is so high that what you end up with is "This company has a beta of 1.3..." (this is the part that gets shown) "... +/- 0.95 given the standard error" (this is the bit that usually gets left out). Sounds like political polls yeah, it's almost always useless to use a regression beta figure. Regression betas are important, but not as literal numbers to use in a valuation. The reason they're useful is that there's a magic theorem in mathematics called the Law of Large Numbers. What it says in the case of betas is that if you take a bunch of noisy beta numbers and average them, then the *average value* you get back is precise. (The larger the sample size, the better the data) Essentially the betas will be noisy in all different directions - some too high, some too low. When you take a pile of them and pull out the average, the errors average out and leave you with a good number. This is a very powerful result in statistics. right! I was a math major for a very brief time and took a couple stats classes in college. (about 17 years ago) Now the point of the beta parameter when used to compute the cost of equity is that it should measure risk. Let me ask another question: risk to whom? Risk to those deviating from the average. ok, what we're talking about is the risk associated with investing in a company. so we're talking about measuring the risk to investors. but what kind of investors? I'm not sure where you're going with this. The obvious thing is that the risk of investing in a compay is the risk taken over investments that are risk-safe or secondarily over just investing in the Indexes. ok, this does get slightly hairy basically in finance there are two kinds of risk associated with a traded company: market risk, and firm-specific risk. Ok. firm-specific risk is what it sounds like, and has the characteristic that it can be diversified away. market risk is the remainder, that portion of company risk which cannot be diversified away. Ok. The assumption in valuation (and finance in general, most of the time) is that the investor we're concerned with is the so-called "marginal investor". So when you invest in an individual company, you are accepting both risks. A marginal investor is an investor that owns a lot of stock, and trades on the stock. Marginal investors are most often institutional investors. Ok The marginal investor is effectively completely diversified. So from the point of view of a Fidelity or a Barclay's or a Vanguard - the investors that actually move markets, unlike you or me - firm-specific risk is not an issue: it has been diversified away by their huge asset portfolios. ok Now make a guess: what percentage of company risk is, on average, firm-specific, and what percent is market risk? 35/65 That guess came from the same place that Dogbert believes sales forecasts comes from. It turns out that on average 80-85% of risk is firm-specific, and only 15-20% is market risk. Doh. :) The reason for belaboring this point is to emphasise what beta does and does not cover. In finance and valuation we look at things from the point-of-view of the marginal investor - and beta measures only that 15-20% of risk that cannot be diversified away. We are essentially sweeping 80-85% of the risk of a company under the carpet and saying "don't worry about that, it can be diversified away so it doesn't matter." Aah, that makes a lot of sense. Oh and yikes! Okay. So beta is our proxy for the risk of a marginal investor in a company. Now we need a way to come up with reasonable values for beta. The approach we'll use is this: we will look at the kind of business the company is in, and use the characteristics of that business to come up with a beta. ok This approach is called a "bottom-up beta" approach, to distinguish it from the regression beta approach. ok There are three factors that determine what value of beta a company should have: (1) how discretionary is their product? (2) what proportion of their costs are fixed? (3) how much debt do they have? ok All other things being equal, a company with a *more discretionary* product should have a higher beta - it will tend to move more sharply up or down relative to the market. Agreed? Definitely. * joshin thinks of LF, TASR, IRBT, and so on. Okay, and (2) and (3) should also be fairly transparent. Now usually these three attributes tend to be reasonably consistent for a given industry. E.g. companies in the "telecom provider" industry sell pretty much the same services, have pretty much the same cost structure, and often (though not always) have roughly similar debt profiles. Ok Telecoms providers and airlines are examples of firms with high fixed costs. These make the profits more volatile, since regardless of how much business is done the costs stay about the same. * joshin fears investing in airlines. as do we all. So with all of the above observations in mind, what we're going to do is come up with a bottom-up beta for the *industry* our company is in. We will do this by *averaging* the regression betas of the companies in this industry, as described. We will also *unlever* the beta we come up with - removing the debt component - and correct it for cash. Ok i'm also listening to this excellent lesson btw Why correct for cash? Well, sometimes you have industries where companies maintain large cash balances. Sometimes the cash balance is 5% or 10% or more of the total value of the company. evn_ - do you have the spreadsheets? if you pasted links, my other computer has them... i'll go back and read it all again later ok jlintz: EMC just had an analyst come in and reiterate a buy rating. Thus the bounce. So if 10% of a telecom provider's market value is actually their cash balance, then in effect the firm is 90% in the telecom provider business and 10% in the cash business. * joshin doesn't understand why people react to analysts so much. Right. In general a given company may be in several businesses, and so their beta will actually be the weighted average of the betas of the sub-businesses. ok Okay, the good news is that we can look up the numbers for the unlevered industry beta corrected for cash - http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/Betas.html Thanks goodness. before we do that Do you think SYY will have a pretty low beta or a pretty high one? * tjl (i=stylus@freenode/helper/stylus) has joined ##finance I would expect their beta to be rather low. Sure, they're in a very non-discretionary industry. :) Yup, staples and food stuffs. What number does the sheet give us? .60 okay, I saw 0.63 AAh, that's right... corrected for cash. Yep. The reason we're unlevering, by the way, is because we will modify our firm's beta with their actual debt profile. ok ok, so we're done with the primary sheet - the Est. 5-yr Growth % is not used in valuation, I just have it there to remind myself what analysts are predicting What did you put in for EST 5-yr growth? Aah, ok. I usually take this number from Reuters - you can find it at the bottom of the "Financial Highlights" page "Est. LT Growth" or something I got 13.48% for SYY Ok. I had 13.4 from either yahoo or google. anyhow, moving on. :) now we need to fill in the secondary sheet, which will give us TTM numbers for things like revenue and operating income. We'll use the 10-K and 10-Q for this. let's fill in the "Year End" column for "This Period" first this comes from the consolidated statements in the 10-K Where? damn uhoh the consolidated statements start on p.34 these bastards don't report their operating income as a separate line item ok, there there are about 2% of companies that do this for some reason so we have to manually subtract COGS and SG&A from revenue to get operating income (also known as EBIT, earnings before interest and taxes) * stylus has quit (Success) Forbes lists it for the last quarter as 454633 whoa there first be very careful about using other pages on the net for financial data - they fuck it up a lot I've found that Reuters' financial statement summaries are usually fairly reliable (but you have to scrutinise them carefully) Yup, just looking for a guestimate to make sure that my calculated numbers are in the ballpark There is an EBIT line in the 10-q I have 35042.08 revenue, 1708.48 EBIT * PhilKC has quit (Client Quit) interest expense 105, capital expenditures 662.56, Depreciation & Amortization 362.56 not sure how you got 1708.48 for EBIT I see 1621.21 These come straight from the statements except for capex, where I've added "additions to PP&E" and "Acquisition of businesses" 35042.075 revenue - (28284.603 cost of sales + 5048.99 operating expenses) = 1708.482 operating income Aah, ok. So ignoring their line item. Since you're pulling out the other numbers Ok. which line item? Earnings before income taxes and cumulative effect of accounting change oh, I see we can't use that number since it's after interest expenses Right, realized that. In FCFF we work with operating cash flows, i.e. before interest expenses, since we will be valuing debt separately. (also operating cash flows are much nicer than cash flows to equity (net income/EPS) ;) (Don't get me started on how incredibly stupid it is to see all of this analyst focus on EPS) (Tell me about operating income or ROC, not EPS which can mean 72 different things, for God's sake) Heh (that should also make you scared shitless to use P/E ratios, since the "E" can mean an absurd number of different things) Yeah, companies have plenty of tricks to manipulate earnings yep. plus there are about a dozen different "kinds" of EPS - GAAP, non-GAAP, pro-forma, fully diluted, partially diluted, non-diluted, and then "forward" versions of all these it's a fucking nightmare Ok, how about depreciation and amortization? taken straight from the cash flow statement The only item there that requires some comment is capital expenditures This is a third area where we deviate from accountants. We want to include everything in capex that really is capex. Usually this is the "Capital Expenditures" line in the cash flow statement, plus several other things. For SYY they've called this "Additions to PP&E" which is also fairly common, but most often you'll just see this listed as "Capital expenditures". The main things we'll add to this are acquisitions and R&D, which are the ultimate capital expenditures. ok Acquisitions are usually reported in the same place in the cash flow statement as capital expenditures, and our spreadsheets will handle the messy business of capitalising R&D. However, please don't be too literal-minded about this - other things could go into capex as well, depending on the company. For example, there's a good case to be made that when valuing a company like Amazon when it was a start-up in the late '90s, a lot of its SG&A could reasonably be considered capex. * joshin googles SG&A (selling, general, and administrative expenses) The important thing is that we want a good and realistic sense of how much the firm is reinvesting back into itself for long-term growth. So R&D and acquisitions should definitely be in there; but in some cases there might be other things as well. I can think of a few companies whose marketing expenses should really be considered capex and not opex, for instance. Yeah, SG&A otherwise known as operating expenses. Ok, spreadsheet looks about done though I have an AVG ROC Divide by 0 error. :) Aha, I still have the Revenue for the Previous period to do. that's the one you need the Q from the previous period for let's double-check our Totals on the secondary sheet: Revenue $9,405.84 $35,042.08 $32,628.44 EBIT $454.63 $1,708.48 $1,495.03 Interest Expense $26.37 $105.00 $109.10 Capital Expenditure $157.29 $662.56 $628.31 Deprec. & Amort. $90.46 $362.56 $345.06 This Period: Revenue 35775.85, EBIT 1770.78, Interest Expense 105.61, CapEx 660.54, D&A 362.96 Prev Period Revenue: 33290.03 (you pasted everything except the Totals ;) Why were your revenues different than that in the sales line item? I'm reporting Totals (green boxes), you're reporting the input values AAh. Sigh, doesn't look like our input values match. let's see...first column looks fine... second one too oh, third one is all messed up your third-column numbers should be taken from the "previous period" column in the latest 10-Q i.e. you fill in the first and third columns at the same time since the info is all there in front of you in the same place Ok. I see. I did it from the 10-k for the previous year rather than the last's year's equivalent quarter. * evn_ has quit () All this secondary sheet does is help us compute trailing-twelve-month values - and since each 10-Q reports the current period alongside the same period last year, we can do it with one 10-K and one 10-Q: fill in the K figures, then add the current-period Q numbers and subtract the prev-period Q numbers. (except we need to pull the 10-Q from four quarters ago to get the "Prev YTD" figure for the "Prev Period" section) Ok, numbers match now. :) ok cool So we're finally done with the input sheet - yay! Let's take a step back and look at the primary page. ok I added a couple of boxes to report some preliminary info based on the entered values - The two most critical numbers for valuation, the Return on Capital and Reinvestment Rate Caveat: These do the right thing with R&D, but not operating leases - I was too lazy for that and it will be taken care of when we copy things into the other spreadsheet ok but this rarely matters much unless you're dealing with a firm where operating leases are significant, which is atypical Anyway, one thing jumps out at me right away for SYY: a return on capital of 21.7%! Better than a stick in the eye. It looks like the average ROC right now for food wholesalers is about 11.5% which is still fairly high actually - I guess food price inflation has been high But 21.7% is a hell of a lot better than the industry average. (do we all know the definition of return on capital by heart and have it tattooed on our foreheads?) Yup. They are big restaurant suppliers and of more than just food. Yup return on capital = (operating income)*(1 - taxrate) / (book value of equity + book value of debt) reinvestment rate = ((capex - depreciation) + change-in-working-capital) / (operating income)*(1 - taxrate) The *growth rate* of a company is the product of its return on capital with its reinvestment rate. This may be the most important principle in valuation. To grow value by a lot, you have to reinvest a lot (high RR) and reinvest well (high ROC). ie... it takes money to make money That's only the "reinvest a lot" half. The other half is the converse of: "A fool and his money are soon parted" This is also where the biggest mistakes get made in valuation. People love to increase value on a spreadsheet by just pumping up the growth rate. But we will always ask: what do you have to do to make that growth happen? It doesn't come for free; it requires reinvestment. We will always see the growth rate as a derived value, not an input. Our inputs are return on capital, and reinvestment rate. I'm going to have to do one of these a day to get this solidly in my head. You also show a 31.35% RR? yup So SYY has a great ROC, but they aren't reinvesting a huge amount - makes sense, since they're a pretty large, stable company. Instead they're paying out a lot of cash in dividends and buybacks. (at least based on TTM numbers alone - a very small view) TTM- trailing 12 months? We won't fill in the Trends sheet now, but basically it's a sheet where you can go back and fill in the critical figures for the last five years. You should do this when you're really valuing a company. yep Finance has a whole new set of TLA's for me to learn. I put the Trends sheet together with the goal of being able to compute 5-year average ROC and RR, which you see on the primary page. Obviously these average values will be undefined/meaningless now since we have not filled in the Trends page. Right, I noticed that when trying to figure out the Div by 0 error However glancing briefly at the ROI and ROE 5-year avg numbers on the Reuters page, it appears that SYY's numbers have been if anything better even than they are now. So probably if we went back and filled in the Trends sheet we'd get similar or better ROC, but possibly significantly different RRs year-to-year. Yeah. RR tends to be pretty volatile, so a lot of times I end up using the 5-year avg RR shown on the primary page as the actual valuation input. ok Anyway, now we're ready to pull out the real spreadsheet. So pull up fcffginzu.xls and head to the Master Inputs sheet... There now. Do you want to capitalize R&D expenses? We'll leave that as No since we don't have any R&D. Do you want to convert operating leases to debt? Yes. Let's hop over to the Operating Lease Converter sheet. ok Here you can fill in our 94.16 value for the current year's expenses and copy-and-paste the remaining six inputs word of caution when copy-and-pasting: Use Paste special... and select "Values" to only paste the values and leave the other cell attributes the same. You should always do this when pasting into this spreadsheet or you'll mess up some cell formatting. Yeah, I've learned that the hard way. :) Once you've done that you should see, lower down, Debt value of leases = 291.75 (this will change later as we fix the other inputs) I don't see that yet. I see 447.73 right now. hmm. I should probably clear the "6 and beyond" field and put in the correct number. Sigh, all better now * joshin goes hunting for caffeine your inputs should be: 94.16 for current year, then 63.38, 53.32, 45.24, 36.20, 27.27, and 142.30. cool when that's done, you can go back to the Master Inputs and fill in the "Inputs from Current Financials" - you can easily paste the main column from the other sheet Should I fill in the EBIT first? Oh, no. It is calculated. you can just paste everything in that column from EBIT down to Value of Non-operating assets Done. then copy the previous-year-end values for revenue and bv of debt/equity Done. ok, now paste the "Current stock price" and "Number of shares outstanding" Done. I jumped ahead. ;) Next box is Market value of debt, which by default is set to the book value of debt you entered earlier Now this is a little complex in principle - it's harder to come up with the market value of a company's debt than of its stock since a lot of its debt isn't traded and you can't just go look it up Ok What we can do is value its debt as one big coupon bond, using the weighted average of its maturities and the firm's cost-of-debt as the discount rate However this is a pain in the arse, and in practice the MV of debt is *very* close to the BV of debt in all but special cases. If you were valuing a fucked-up company like GM you would want to be careful about the MV of its debt. But 99% of the time we can safely use the book value as a proxy. Yeah. Ok. so we'll leave that one as-is at $1976.08 Would you like to use the book value debt ratio? leave this at No, and we'll come back and set the box below, the debt ratio to use, in a little bit. ok Now the General Market Data section. We need the current risk-free rate, and the 'risk premium for equity' I mentioned earlier. Remember how to find the risk-free rate? 10 year treasury rate strictly speaking, the 10-year yield on a *zero coupon" bond So around 3.65% http://online.wsj.com/mdc/public/page/2_3020-tstrips.html you can get it from the Stripped Principal section for Treasury Notes, last item in the last (Nov. 2017) in the list, even Ok. which is 3.69, slightly higher than the yield for a normal 10-year T-note, unsurprisingly * stylus (i=stylus@freenode/helper/stylus) has joined ##finance Now for the equity risk premium, let's use 4%. This is a complicated topic, but briefly 4% is an average value for the implied premium reported by the U.S. market over time Jeez, there's be a market for a nice consolidated site with all this info. ok it appears that, in general and on average, investors in equity markets in developed countries demand about a 4% premium for putting their money in equities. if we were valuing an emerging market company, we would need to adjust the risk premium for country risk and it would get more involved. ok (I am assuming SYY does not get a large chunk of its revenues from other countries) Nope maybe they do, I don't know. :) if so, then 4% could well be too low of an equity risk premium. Depends on if you consider Canada to be another country. ;) They have a presence there. okay, next is the Ratings section. we need to find out what the firm's cost of debt is, i.e. what is the rate that they can borrow money, long term, today. 24/177 of their facilities are in Canada * Aiken (n=james@ppp59-167-109-167.lns3.bne1.internode.on.net) has joined ##finance The easiest way to do this is to look it up. If you go to www.standardandpoors.com you can do a Ratings Search and get the company's credit rating, if they're rated by S&P. yeah, Canada will have the same equity risk premium as the US; same with Britain, etc. I just looked up the rating for Sysco and got AA- ok. I had to register there before I could do a search if they weren't rated, we could estimate their rating using the Ratings Estimator sheet yeah, you have to register, but it's worth doing so you can look this number up for valuation :) Yup and none. /none/done Anyway, we want to glance at the Ratings Estimator sheet anyway because it has a mapping of ratings to spreads In this case there isn't an AA- but we'll be a little kinder to Sysco and use AA, which shows up as a 1.25% spread * tjl has quit (Success) So going back to the Ratings section of Master Inputs, we'll leave "No" for Do you want to estimate the firm's synthetic rating (since we know what the real rating is), and then enter 1, AA, and 4.94% in the next three boxes ok. (where we've added the spread of 1.25% to the risk-free rate to come up with SYY's cost of debt) Yup. ok, next is Options, but we'll put No here. We could use this section to account for the value dilution of options, but I think option dilution is unlikely to be a big issue with a company like SYY. if we were valuing Amazon or Yahoo it would be a big issue. Yeah. How would you calculate that for them? This is also complicated somewhat by the recent regulations about expensing of stock-based compensation. Basically you have to find out how many options were granted before the new rules came into effect and value those in the Options section, but omit any options granted after that point because they're already built into the cash flows as operating expenses. well, the Option Value sheet does that for us - we just need to fill in the inputs in the Options section if we want to do it Bah, sold that EMC a bit too early. Didn't want to hold it but figured it was good for a bounce and the in-money call options were good even if it dropped. The numbers are easily found? everything is in the 10-K except for an updated Standard Deviation of Stock Price which you can actually get from Morningstar's options quote page. (Historical volatility, I think they call it there) Nice. anyway, we'll put No there and move on to the fun part: the Valuation Inputs section. ok Now a couple of comments about what we've already done and what we're going to do. Everything we've input so far is just historical data; nothing controversial there. The hard part in valuation is making future projections. We need to project the firm's cash flows out into the future in a sensible way, and also adjust for changes that happens when it becomes a "stable growth" company. Ok. SYY is probably already a stable growth company. Well, that's an interesting point. If our company already *is* in stable growth, our valuation model becomes much simpler: it's a one-stage model, where the only thing we need to compute is a terminal value. Remember how we said valuing a company is like valuing a bond, except the finite sequence of future cash flows becomes an infinite series? In order to make that series converge, we must stop at a certain number of years in the future, and declare that from that point forward, the firm will grow at a constant rate forever. Ok This "constant growth rate forever" piece is what's used to come up with what's called the "terminal value" for the firm. Understood However there are some severe restrictions on how a company in stable growth can look. The biggest one relates to the growth rate. There is a cap which the stable growth rate of a company cannot exceed. Any idea what that might be? Not sure. I'd guess it would be around 8%. Any particular reason for the magic number 8? :) its near the long term riskfree rate + risk premium for equity One way to think of it is that the stable (infinite!) growth rate for a company can't exceed the GDP - if it did it would eventually become the GDP. Or I could go with the historical inflation rate plus the long term risk free rate Aah. Furthermore, we have a very good proxy for nominal GDP growth in the form of the risk-free rate, which captures exactly "real expected GDP growth rate + expected inflation". So the cap on stable growth is actually the risk-free rate. Interesting. I was wondering when inflation would pop its head in. With a risk-free rate of 3.69%, we can look at SYY's current growth - as we remarked earlier the consensus has it around 13.5%, which is quite a bit higher than 3.69%. So at least based on that, it seems SYY still has some "more than stable" growth left. Yup We could be a little more thorough and go and actually look at its growth rates for the past few years to get a better idea... Reuters has some info under Financial Highlights - e.g. its Sales Growth % over the last 5 years averages 8.46% (Sales is better than equity earnings for us since we're doing FCFF, so that analyst number, apart from being dodgy just because it's an analyst number, is of the wrong type) Ok We could further go back and look at operating income growth for the last couple of years, etc. But I think it's safe to say that it's still growing a bit faster than 3.69% - good news for investors! :) So what we're talking about here is input # 54 - Length of high-growth period. I recommend you use one of three values for this input: 0, 5, or 10. Use 0 if the company is already in stable growth. Then everything in this section is moot. Ok. The company is looking at growth accelerating, at least for the short term so 10 might make more sense than 5. Your other options are 5 or 10, which you will have to choose based on your assessment of the company. Bear in mind that *very* few companies ever achieve > 10 years of high growth; in fact, you can probably name many of the few that have. If that 13.4% is close Ok, never mind. 5 it is. I should say "> 10 years of high growth relative to their industry" because the other possibility is that the industry itself is growing - this can be an important factor too. true. So we really don't use numbers > 10, because it rarely happens and also it just gets too hard to project cash flows out that far. Why not numbers like 3 or 7 or 9? Well, if you think you can predict exactly 7 years of high-growth cash flows, more power to you - but my information never seems to be precise enough for that. :) :) Let me make a suggestion here which is the rule I tend to use more often than not: If it's a good company with reasonably strong growth prospects given the business it's in, you can use 10 here and then select Yes in the input later on which asks, "Do you want me to gradually adjust your high-growth inputs to stable growth levels?" What this does is, it uses your high-growth values for 5 years, and then in the second 5 years adjusts things like return on capital down gradually to the stable growth level. That makes sense. This gives us a 3-stage model: 5 years high growth, followed by 5 years of diminishing high growth, followed by stable growth forever. You can use this sheet to get either 1, 2, or 3 stage models. 1-stage is when it's stable growth. 2-stage might be, you enter 5 here and then "No" in the box mentioned, so you model a straight 5 years of high growth and then stable growth thereafter. Ok, so what are your thoughts on SYY? You could even decide, hey, I think this company is going to go for 10 years at high-growth levels, so I'll use 10 years high-growth and not adjust the last 5 years down to stable. Yeah but I'd rather not. I have a tendency of overvaluing companies. ;) Well, I should caveat any thoughts here by saying I know fuck-all about SYY, not having done any research at all. And this section is where your research pays off. This is where you go off and find out all you can about the company and get an idea of its prospects; then you use the model to put your assessment into numbers. I haven't paid much attention to them lately. I just saw one of their trucks yesterday when driving home and thought I should see how they're doing. However, since we don't have that research to hand, we'll make some reasonable guesses based on what we know. Do you have SYY shares? Not currently. I've traded in and out of them for years. Been about a year since I last owned them. ok, cool. Well, based on their impressive return on capital, let's assume they still have some fairly good growth left. We'll use a 10-year, three-stage model. ok Next one: Beta to use during high growth. Now to get this number, we need to take our unlevered industry beta and lever it up based on SYY's debt ratio. We can do this on the Bottom-up Beta sheet. ok (which gives me 0.67) ditto next, Do you want to use the debt ratio computed from your inputs? We'll say yes, since we don't have any other info suggesting a different debt ratio would be better. heh (For instance, we might know that they're about to take on another $10b of debt based on a press release) If so the debt ratio that will be used to compute the cost of capital will be: 11.68%. Ok Now let's fill in a gap from earlier: take that 11.68% and put it into input # 32. Next: Do you want to keep the existing ratio of working capital to revenue? Where did you get the 11.68%? Isn't that what the sheet gives you in B57? ah... yes it does ok A comment about #59: remember we said that "change in working capital" is an input to computing the firm's reinvestment rate. However this number is very volatile, so rather than using whatever the latest change-in-wc is based on the most recent financials, it tends to work out better to regard WC as a fixed percentage of revenue. alright In most businesses WC tends to track revenue pretty well. So #59 asks if we want to use the current ratio of WC to revenue or some other number. Now one of the things my Trends page computes is the ratio of WC to revenue. So if you filled in the Trend sheet, you would be able to look at the last five years and see what the WC/Revenue was. ok Then you'd know if the current figure is anomalous or pretty much on track. But since we don't have that filled out, we'll assume the current one is good. Actually, we do have both this year's and last year's WC and Revenue, so we can at least look at both of those. And the last one was 4.28% compared to the 4.11% for this year, shown in #60. So it looks like 4-4.5% might be about right for SYY's WC/Revenue ratio. ok So let's just say Yes for #59, and then #60 reminds us what the current WC/Revenue is. Next: Do you want to compute your growth rate from fundamentals? And the answer to this is always: yes! heh Now #64-#66 says: If yes, the inputs to the fundamental growth calculation are: Return on Capital: 20.65% Reinvestment Rate: 36.22% yup (Notice these differ slightly from the ones on my other spreadsheet - that's because of a few small differences, like my not accounting for operating leases) right BTW, what did yout put in for 60? 60 is actually an output, not an input it's computed at 4.11% if we wanted to change it, we would have said No in 59 and then put something else in 61 ok Now in 67-69, we have the opportunity to change the values of ROC and RR. We'll say no here, since we don't have better values, but again if we had filled in the Trends sheet, we might for example decide to say Yes here and put in an Average RR and/or Average ROC instead The important thing is that we want our ROC and RR values to be the best, most likely figures we expect the company to see over the next 5-10 years. Maybe that means the figures from the most recent TTM, maybe it means averages over the last 5 years, maybe it means something else altogether. Ok. These are the two most critical variables in the valuation, so this is where your understanding of the company, its prospects, and its competetive environment really comes into play. Makes sense. Now for 71, we'll say Yes, since we agreed to use a 3-stage model. ok Last section...almost done! Stable growth figures. What will this company look like 10 years down the line when it's a stable growth company? First, what's the growth rate? Remember it has to be < the risk-free rate. Useful piece of information: it actually matters very little what you put in here. Good because that risk-free rate will likely be pretty different then. Because we compute growth as ROC*RR, what we do in the stable growth portion of the model is take the growth rate that's entered and back out the reinvestment rate required to achieve that growth. What this means in practice is that any change you make the growth rate will be offset by higher reinvestment, so the net effect on the company value will be almost zero. ok Anyway, let's go ahead and use 3% as SYY's stable growth rate. Now for the stable-growth beta, we want to adjust the firm's beta toward 1. The rule we use here is that betas of stable companies lie between 0.8 and 1.2. Since SYY has a low beta, we'll move it up to the minimum, and give it a stable-growth beta of 0.8. ok For 76, risk premium for equity, this is set to our earlier input of 4% and we can leave it that way. Now for the debt ratio in stable growth, a good thing to do here is compare the firm's debt ratio to the average for its industry. I'd think the debt ratio would go down as they'd not be financing growth The debt ratio question is actually one of capital structure, more than growth as such. This is a corporate finance topic, but basically the firm will need to finance itself with a combination of equity and debt, regardless of its growth rate. Different companies and industries have different optimal debt ratios - one of main jobs of a good CFO or corporate finance consultant is to help the company achieve the debt ratio that minimises its cost of capital. Right. You already covered the best ways to raise capital. Now we can look up the industry figures - http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/dbtfund.htm we're interested in the 3rd column, MV debt ratio 32.57% Yep, so as a rule this industry has a debt ratio that's significantly higher than SYY's current debt ratio. (I noticed earlier when we were gathering the numbers that it had a fairly small amount of debt relative to its size) ok There are several reasons why this might be, but it's interesting to note that SYY may need to take on more debt in order to come closer to its optimal debt ratio. Its cost of debt, as usual, is significantly lower than its cost of equity, so changing its debt ratio (say by borrowing money and using it to buy back stock) could lower its cost of capital. Yup. Let's assume that 10 years down the line it's going to be close to the average for its industry as far as the debt ratio goes. So line 77 should be 32.57% yep Now for 78, we know what its current cost of debt is based on its rating. Usually in stable growth a firm will have a lower cost of debt since it has more cash coming in, it's more stable, etc. But you rounded up to AA from AA- so we could probably leave it there. Yep, AA is already a pretty good rating; not that many companies really have AAA. So we'll be a little conservative and leave that one alone. For the tax rate, we could use the marginal rate of 35% but it seems to have been paying more like 38% for quite a while, so let's re-use the 38.24% tax rate. For 81, Do you want to compute reinvestment needs in stable growth based on fundamentals? Again we should usually answer Yes here. Ok And now the most significant variable for stable growth, and our final input: If yes, enter the return on capital the firm will have in stable growth. There are a couple of choices here. The thing to realise is that if you input a value here that is larger than the firm's cost of capital in stable-growth, you are assuming that the firm will make excess returns in perpetuity. Which would be nice but isn't terribly realistic. This is a very strong assumption. In principle no one ever makes excess returns in perpetuity, so the conservative thing to do here is to set the stable-growth return on capital equal to the stable-growth cost of capital. ok However, this is where our models and mathematics don't always fit perfectly with the real world. It's also somewhat unreasonable to say that a firm with very high excess returns will suddenly lose them all after exactly 10 years. true What I recommend is, if you're valuing a company that you don't think will necessarily retain any strong competetive advantages 10+ years out (which, of course, means most companies), then set the return on capital in stable growth equal to the stable-growth cast of capital. If you are valuing a company that you expect to have competetive advantages 10+ years out, you can set this to a couple of percentage points higher than the stable-growth cost of capital. Probably not more than 3-4% at the extreme. The "Cost of Capital in Stable Phase" appears on the Valuation Model sheet as D55. So 5.7 + [0-4%] depending Yeah, I don't know that much about SYY and especially given the nature of the industry I'm inclined to be a bit conservative here. Let's set this to 6% for now, just a bit above SYY's stable-growth cost of capital. ok That's pretty conservative. :) Yeah, it actually is considering how well they're doing now. Okay, we're done! Let's look at the Valuation Model sheet. Starting at the top, this sheet has the outputs to all the critical figures computed from our inputs. Very cool. We have the adjusted operating income, capex, etc. based on R&D and operating leases, the computed growth rates for the high growth period and for stable growth, costs of equity, debt and capital in high-growth and stable phase, year-by-year cash flow projections, etc. According to the chart, SYY is significantly undervalued, even with being as conservative on the ROC in stable growth Thanks so much for the time in going through this with me. Yeah, so the big number at the bottom is always fun to check: instant feedback vs. the market. :) :) $44.14/share vs. $28.17 at today's close so there are several caveats to take on board here about this valuation One is that we basically did no research to come up with our estimates. I picked what I think are "reasonable" values, but I might have changed any of them significantly if I had gone out digging around on the company. Past performance is not indicative of future results. Maybe their return on capital is going to get cut in half starting next year, in which case our growth figures would be off by a lot. Or maybe they're going to reinvest 80% instead of 30% and actually grow a lot more. What I like to do is put together a "first approximation" valuation like we just did, and then examine the key assumptions against what I know about the company. It's especially useful to find out what the key variables are that drive the valuation, and how they might differ from expectations. Yeah. How well has this method worked for you? It certainly seems to have a better base in reality than TA or social analysis. /seems// Usually the most important variables are the return on capital and reinvestment rate. By making assumptions about these you are effectively quantifying the performance of the industry and the capability of the company within the industry to generate excess returns (ROC), and how much the company's management will be putting back into the company to fuel growth (RR) vs. paying out the money in dividends or buybacks, or just ... ... accruing an ever-larger cash balance. While this method is a lot more work than how I normally value stocks, it seems a lot less haphazard. Well, as I say intrinsic valuation is the backbone of valuation, and understanding how it works is the financial equivalent of anatomy. You can use other methods, but you will not use them properly unless you understand DCF valuation. This is how (good) investment banks, for instance, actually value companies for M&A. (except 7 times out of 10 they fuck it up royally - there are some amazing stories) Yeah though still better than how well the tech industry does it. From a personal investing approach, it works well if you understand what it is as a tool and how it fits in to the picture, (e.g. use it as a tool to clarify and quantify your projections about the future), and if you are disciplined in doing valuations and updating them before making decisions. When valuing any company, there are three possibilities: (1) the market is right and you're wrong; (2) you're right and the market is wrong; (3) both you and the market are wrong. Yeah. My methods have generally worked well when I am disciplined but they're not as well defined as what you just taught me. To make money on a valuation, you have to (a) be right where the market is wrong, and (b) the market has to come around to your point of view. Now I just need to figure out how to automate this a bit more. The market is eventually rational. It doesn't do any good to buy an undervalued stock if the market never sees the error of its ways and corrects back up to fair value. :) however, as an individual investor you have one huge advantage over the institutions, which is your time horizon. Very true. Markets can be way out of whack in the short term, but in the long term they tend to get things mostly right. Institutional investors usually have a time horizon of < 12 months - but it often takes the market a lot longer than that to correct. Yup. As individuals we can value a company like SYY and say, you know, this stock is being unfairly punished because of such-and-such, and may continue to be for another year, but in 3 or 5 years I think it's still going to be going strong, and by then the market will be running to catch up. Yeah. Yup. And in the case of SYY, you'd have been collecting a decent dividend in the mean time. (As well as possibly selling covered calls at/near the intrinsic valuation point) This really helps me at the part of investing that I'm worst at: Knowing when to sell. I guess they actually just released Q2 earnings yesterday so we could have used those numbers; just a matter of updating the secondary sheet noo broker sent tax papers Their operating income grew 10.9% in Q2 vs. last year, which is actually well ahead of our expected 7.48% growth rate when people say growth rate what is it the growth rate of operation income? in this case, it's operating income. well, usually Heh, yeah. Saw that when we were around 2/3rds through the example. Usually when you hear it, it's a very good idea to ask that question. cause operation income seems like a bad one to look at by itself to decide growth Most often "growth rate" will refer to EPS or net income. How can you "decide growth" from looking at any bare number on the income statement? by the way, once you've done an intrinsic valuation like this, you can derive exact numbers for what the correct ratios like P/E should be when i am given a problem (so far) to calculate a future value, i am already given the growth rate i do that in my sleep by now lol Yeah, well, you've been given all the hard work then - the rest is just computation. ;) km = kRF - blah blah i know well this semester ill learn all the details g2g Another amusing thing we can do is play with the variables and see what we'd have to assume for our value to match the market's value. yup, was doing that