Friday, March 22, 2013

Valuing Howden Joinery Group



My posts on Howdens Joinery tried to understand its business model: how and why does it make money? That must be the very first question that a prospective investor asks of any business that s/he invests in. The arithmetic of valuation is the second step, and I’ll turn to that now.

A Rough Cut:

Howdens was once twinned with a low quality furniture retail business. Howdens sold that business to a private equity firm in 2006 and, when that business went bust in 2008, the legacy store lease obligations of the bankrupt business were, for one reason or another, put to Howdens.  Since 2008, therefore, Howdens had been both paying rents on properties it doesn’t use and paying breakage costs in order to rid itself of these legacy properties. These extraordinary payments are now almost entirely in the past; annual payments of only £2 million remain.


Also, Howdens’ pension fund assets once enjoyed assumptions that perhaps would have seemed reasonable at the time: an 8.4% rate of return on equities and a 4.5% rate of return on government bonds. As these assumptions crumbled in the face of the downturn, the net pension liability ballooned, and the trustee insisted on large cash infusions to eliminate the deficit. As it stands today, the deficit amounts to 154 million, the weighted average expected return on pension assets has contracted from 7.39% to 5.07% (Equities: 6.25%; Bonds: 2.55%) and Howdens has agreed to inject a further 45 million a year for the next three years. By 2015, the net liability should be eliminated: I doubt that there’s a reasonable basis for suggesting that expected long term returns on equities or bonds ought to be lower than they are now  and the cash payments will therefore be enough to make the retirement plan whole.  

We’re ready, now, to make a rough estimate of the lower bound of Howdens’ ability to generate free cash. We add back the legacy rent payments, the lease breakage costs, the cash contributions to the pension fund, and we are left with 23p in free cash per share, or, at the current market capitalization, an 8% FCF yield.



This is a rough cut and it’s wrong: we know that Howdens is growing, so that free cash flow doesn’t represent earnings; and we know that the years 2008-212 are exceptional ones, such that earnings in these years understate the company’s long-term earnings capacity.

So what is Howdens’ true earning power?


Approximately Right, from the Bottom Up

Let’s start with sales. We know that Howdens’ sales per square foot is a function of cyclicality and store maturation. The combined effect of these two factors can be seen below:


Average sales per square foot between the top and bottom of the cycle, i.e. between 2005 and 2012, is £199.4.

We know also that the depots mature with time. We can isolate this effect by calculating the CAGR in sales per square foot between the two bottoms of the cycle, 2002 and 2012. It’s not a perfect approach – that would take more information that we have available to us – but it’s good enough. The maturation effect is 1.4% per year.

Put these two factors together and look out to 2016 and we can estimate that sales will amount to something like £1,475 million: 700 stores, 10,000 sq ft per store, £211 in sales per sq ft.

We know that adjusted operating margin is 17.45%, that depot rents approximate £5 per square foot, that the business needs 30 days of noncash working capital, and that it costs £170,000 to build out each new depot so, at this point, we can estimate earnings and cash flow:




The business will generate about £650 million in cash over the next four years. Of this, 180 million will be contributed to the pension plan, £46 million will be tied up in increased working capital, and £30 million will represent cash capex for 171 new depots.

This leaves £390 million in free cash which, for the sake of simplicity, I choose to keep on the balance sheet. In reality, some of it will be paid out in dividends but it makes little difference to the valuation whether it is or isn’t.

A 2016 equity valuation of £2,775 sees the shares being worth 655p each, representing a return of 29% a year.

I’ll pause here to re-emphasize a couple of points:

·      This estimate of Howdens’ intrinsic value would be impossible without confidence in the robustness of Howdens’ business model.  One can’t do this for some teen fashion retailer or low-grade home supplies distributor and reasonably expect to be close: tastes change, and anything that can be disintermediated will be, sooner or later; forecasting revenues and profits for these kinds of businesses is, in my view, a fantasy.

·       The above valuation is dependent on two assumptions only, neither of which is fanciful: (a) things stay as they have been; and (b) management’s 700 depot target is proved reasonable.     

      If you'd like to take a stab at breaking my valuation, I’d welcome it in the comments below.


13 comments:

  1. Hi

    Have read your posts on Howden with interest. How did you come up with enterprise value of £2,575 in 2016? Are you assuming some premium to operating assets (ie a price/book ratio with some growth thrown in)?

    Thanks

    ReplyDelete
    Replies
    1. Hi,

      I think 700 depots gets us to 206 million in annual profit. I've capitalized that amount at 8% to get 2575 in enterprise value. I think 8% is a reasonable estimate of HWDN's long run cost of capital.

      Delete
  2. red

    I like your approach, but I have a few attempts at breaking your valuation.

    1. Your assumption of achieving 700 stores by 2016 seems aggressive as that implies over 40 depots/yr. HWDN hasn't opened this many depots since 2007 and management has targeted 20-30 the past few years. Perhaps I missed something where they upped their targets, but if not 2016 seems too soon for the 700 store target.

    2. Your assumed sales/sq ft seems aggressive as I wouldn't call 2012 a bottom of a cycle sales level, rather more of a trend in store saturation. Management has even mentioned that opening more than one store in a town (in-filling) results in lower sales per store (but better service which is probably better in the long-run). Additionally, if HWDN were to accelerate store openings as you assume it would negatively impact sales/sq ft for at least the next three or four years (using management's stated 7-yr maturity cycle for stores) as a relatively large number of new stores would be driving sales growth.

    3. While management states depot fit-out is around £170k, cap ex numbers indicate actual spend on depots is closer to £400-450k (£9m in 2012, £8m in 2011 for 20 new depots each year). This includes expansions and relocations, but one would presume these would continue in coming years. This would more than double your resulting cap ex spending on 171 new depots.

    If you take a more conservative view and assume it takes until 2018 to achieve 700 stores (30 per year) and sales/sq ft rise from £170 next year to £183 in 2018 and use your margins and tax rate it looks like NOPAT in 2018 is closer to £170m and value of the business would be £2.1bn. If I adjust your cash flow numbers for the higher depot costs cash would be about £675m by 2018 and the implied value of the equity would be £2.5bn, offering a 10% annual return from current prices.

    Anywho, that's my attempt to break your valuation. Let me know where I've overlooked items and gone astray (as I'm sure I have).

    ReplyDelete
    Replies
    1. https://docs.google.com/spreadsheet/ccc?key=0AoCLhoPnjQDgdFdXemFiUWxHV1VTdDZYT2doM0E1bUE#gid=0

      Let me start here: one can see in figure 1 that in the period 2006-2012, the number of depots rose by 37% while EBITDA (ex legacy expenses) rose by 91%. This is the store maturation effect but we’ll put that aside for the moment.

      The most serious/material question is about capex. Table 2 shows that the increase in gross PP&E since 2008 is 30 million. We know that 20 million was spent on a new manufacturing facility in Howden & Runcorn. 10 million remains. Divided by the increase in depots, 30, gives capex of 200,000 per depot. There is no doubt some expenditure on distribution systems and central infrastructure that I haven’t accounted for so that the guidance of 170,000 per depot seems plausible.

      Sales per sq foot as a result of economic reflation: it is possible, though unlikely, that the effects of Osborne’s budget will bypass HWDN and that, in any case, sales per square foot never reach the levels attained in the past. I have used your sales/Sq ft forward estimates in Figure 4. What I have resisted is your assumption about margins: a marginal increase in sales causes a marginal increase in COGS but not in fixed store costs like rent, labor, electricity, central overhead burden etc. There is operating leverage and increased gross profit falls more or less intact to the pretax profit line.

      So, with sales per square foot that never again achieve 2007 levels, with no store maturation because of the deleterious of geographic infills, and with store growth at 30 per year, the equity is worth 313p. This is the margin of safety price: everything that can go wrong, does.

      Delete
    2. I've begun looking into Howden's and just requested access to that google spreadsheet. Would be grateful to see it if you're willing. Thanks, regardless, for all that you share on this blog.

      Delete
    3. Done. Thanks for reading.

      Delete
    4. red

      Thanks for the detailed reply and we may have to agree to disagree on increasing sales intensity.

      However, I think you've pulled the wrong numbers for Gross PPE as I've got 2012 being 316m and 2008 being 266m for a 50m difference less £20 on H&R giving us 30m remaining across 75 new depots for a cost of £400k per.

      And point taken on the operating leverage and I agree in theory, but historically selling and distribution costs (the largest component of operating expenses) hasn't scaled at all - actually having ramped up as a % of revenue from 31.5% in 2007 to 40.5% in 2012. An argument could be made that this is a result of new, immature stores, but I'm not sure I'm willing to give that to management without some proof.

      I'd also be cautious in saying this is an "everything that can go wrong" scenario.

      Delete
    5. Thanks for pointing out the mistake in 2012 Gross PP&E. I need to fix that.

      Selling & Distrib is a bit of a catchall category: take out the legacy rent and lease breakage charges and selling admin is steady at ~340K per store.

      There is no protection from the unknown unknowns for anyone. I try to avoid the known unknowns. If I know how and why a business works, however, I use the "beyond a reasonable doubt" standard.

      Thanks for the follow up.

      Delete
    6. red

      I'm sorry to be a pest, but I just can't get my numbers to match yours. If I back out legacy rent charges from Selling & Dist costs (16.4 in 2012, 29.4 in 2011, 37.5 in 2010, 46.8 in 2009, and 46.5 in 2008 - and I think you've double counted the legacy rents in your 2008 adjustment) I still see them increasing as a % of revenue from 31.3% in 2008 (251.8/805.7) to 38.1% in 2012 (337.7/887.1 - I adjusted the 2012 number for the £5m in extra costs from an extra day of trading, too).

      Management's comments would indicate wage inflation being a big part of this.

      Additionally, if I look at S&D as averaged across stores I get £555k in 2008 (251.8/454) rising to £638k in 2012 (337.7/529) or 3.5% inflation per year. You and I may be looking at a different set of numbers, which is what I'm trying to clear up here.

      Thanks for your patience.

      Delete
    7. I've had a fresh look at it:
      https://docs.google.com/spreadsheet/ccc?key=0AoCLhoPnjQDgdFByRVp4WFNONzFONDJJVVhzWmYtaHc#gid=4

      I think that costs/revenue are likely to rise because sales/depot has fallen off.

      But one would expect that S&D costs per depot (and ex-COGS total operating costs per depot) should be falling in real terms, whatever one does with the legacy rent charges.

      So, I'm seeing reported selling & distribution costs rising at half the UK core inflation rate.

      I have used the "continuing operations" figure in 2008 and that may be another source of the difference in our results.

      Thanks for the questions -- I appreciate the opportunity to revisit and make sure that I haven't made serious mistakes.

      Delete
    8. red

      Thank you for that explanation. I would ask, however, if you think looking at 2008 to 2012 is a representative example as it would appear 2008 was an odd year (with a lot of one-time expenses and disruptions to the company). If you move your analysis to the right one year inflation jumps dramatically.

      I guess this investment rests on if HWDN can get its sales per sq ft back up and actually see its operational leverage come through. I don't think we've seen that to date, but it is possible.

      Delete
  3. Hi, are you still following this company?

    ReplyDelete
  4. It would be a value purchase at under 250p

    ReplyDelete