Saturday, 16 April 2016

Gluskin Sheff (again)

After I wrote my initial piece on Gluskin Sheff, I got to thinking, apart from the obvious "$185M" reason for current under-performance while the rest of the market has worked itself into a frenzy, is there anything else I'm missing here?

Here's a 7 year weekly study comparing GS vs. the Dow Jones US Asset Managers Index.  Up until February, there certainly appeared to be a close correlation between the %ge returns between GS and a proxy for US asset managers:

And here's the same study over just the last year:

Fairly closely correlated from my point of view with the exception of the last couple of months which I'll attribute to concern over how much the eventual settlement will end up costing.  Which begs the question, why the overall weakness in not just GS but all US Asset Managers over the last year while the broader market (measured by the S&P 500) is knocking on heaven's door?

Believe it or not, I don't have the answer (no one does).  I can only observe that for now, while the broader market as measured by the S&P500 sits about 2% removed from eclipsing it's all time highs made in the summer of 2015, the US asset manager index is 17% below it's 2015 high, and GS is almost 45% below it's 2014 high!

And I begin to wonder whether asset managers are a leading or a lagging indicator.  Surely, if the S&P500 takes it's 2015 highs out, shouldn't asset managers on balance, benefit?  Shouldn't exposure to asset managers be a logical consequence of a market making (or approaching) new highs as they theoretically benefit from fresh fund flows looking to buy (or chase) new highs?  Again, I don't have an answer.

All I can do is try and focus on determining whether GS makes sense as a possible candidate, or not.  So, on this note, here are some further observations on GS.

On the subject of AUM, I performed the following analysis vs. both management fees and performance fees to get a feel for how lumpy overall fees are, and to get a feel for how AUM have grown.

Overall observations:

  • AUM have grown, by virtue of both regular fund inflows, AUM appreciation during a generally uninterrupted period of rising markets and by virtue of acquisition (GS acquired Blair Franklin in 2014)
  • Base management fees seem like the bread and butter of the business, but by no means can they be counted on for growth.  As a %ge of AUM, base management fees have come down from 1.47% in 2008 to 1.24% last year.  Given the proliferation of low fee ETF's, my guess is that base management fees will continue to come under pressure.
  • The real icing on the cake are the performance fees, but, you can get a feel for how lumpy performance fees are.  In some years, GS does really well (see 2014), other years, not so much (2012, 2009).  I wonder if 2014 was an anomaly type year as the huge bump in performance fees correlated pretty closely with GS' all time high (it's been downhill ever since).  So my question...what the heck happened in 2014 to juice performance fees, and is it repeatable (a point for further research, below).

Here's what scares me about a retail nobody, I have no idea what they're doing in terms of their AUM program internally.  I can get an overall picture of the distribution of AUM by program from the AIF (see below), but at first glance I really don't know what they're doing within each program.  I note that around 30% of overall AUM are run through fixed income / alternative strategy programs, which makes me start wondering about potential leverage employed concomitant with running a hedge fund (or funds).  And if hedge fund strategies are employed, does this make GS more or less comparable to (or riskier than) other Canadian Asset Managers?  The correlation of price performance since the 2014 highs with the drop off in performance fees since 2014 could be symptomatic of investors not believing that 2014 is repeatable, but, I also note that part of the push up to $30+ in 2014 was likely due to speculation that GS was on the sale block.

I believe that any investment thesis for owning GS must be predicated on the belief (or leap of faith) that 2014 was not an anomaly year in terms of performance and that such potential future performance is being under-priced by the market and is being obscured by the current worries over the eventual settlement amount.

So here's what happened in 2014:


And here's what happened in 2015:

And here's what's happened in 2016 through Dec 2015:

From the 2014 annual report re: 2014 performance:

And from the 2015 annual report re: 2015 performance:

So overall, pretty vague comments about specific drivers of performance, but from comparing 2015 vs. 2014, it sounded as as if they switched gears from Canadian exposure to bump US and International exposure, while remaining cautious.

Concluding Thoughts

Personally, I think that any attempt at valuation here has to conservatively discount performance fees as there's no way of knowing when (or if) they're going to repeat 2014.  I've already attempted one method of valuation previously on an expected free cash flow basis.

I suppose another way of getting a handle on valuation is to treat current AUM x average fees excluding 2014 of say, 1.9% and treat the cash flows as a perpetuity with no growth.

In this case, quickly, I get $8.5B x 1.9% = $160M, less salaries, G&A & occupancy (assume 55% of $160M) = $160M - $85M = $75M (pretty close to my free cash flow estimate), and discount this perpetually at say, 10%:

$75M / .1 = $750M

At 12%:

$75M / .12 = $625M

At 15%:

$75M / .15 = $500M

Current EV = $460M, so even in the worst case, with a 15% discount rate, my rough estimate of perpetual no growth all in fees produces a theoretical $500M value compared to $460M currently. This is before any bump in performance fees due to outperformance, and not counting any possible premium attributable to GS if it became a takeover target.

The caveat here is that AUM don't drop precipitously due to overall market weakness.  I note that in 2008/2009, AUM's dropped 20%, so GS certainly wasn't immune to the effects of the crisis, even if the drop wasn't attributable to actual redemptions.

No comments:

Post a Comment