Thread by @bauhiniacapital

@therobotjames For the sake of playing devil's advocate, I will take the other side of that, "for most" for a couple of reasons.

I will start with the following assumptions (which may be wrong-footed, but their mine):
a) there is no tax differential between high turnover and low turnover

@therobotjames b) we are talking institutional - comms are 2-5bp, cost structure in mgmt fees is, say 15-20bp/yr
c) end investors WANT active share risk which is why they pay you the big bucks to provide active risk.

@therobotjames What IS certain in this case is that
a) you are in the active mgmt business so it is CERTAIN you have inherent expectations of creating positive alpha vs benchmark. You may not succeed every week, but on average, that's your business model.
b) Moving from say 4.5bp/yr comm

@therobotjames friction to 9bp/yr (avg comm rate 3bp, moving from 18mo avg hold to 9mo), means your particular alpha bogey has gone up by 4.5bp/yr.
c) everyone on twitter is a star so can easily outperform benchmarks by 10% a quarter, and is a stockpicker/timer extraordinaire so adds 5% to the

@therobotjames value of a stock just because they decide to look at it.

I am being slightly facetious with 'c' but expectations change vs ex-ante when you spend time on something (even if the company didn't know they were supposed to outperform more because Chad VestBro Capital made a model).

@therobotjames If we assume that average hold halves from 9mo to 4.5mo, then the are now up to 16bp/yr in comm friction (tho maybe less as you negotiate lower comms, or just use a different broker).

The trick?

IF one is performing well - outperforming the alpha slope one had modelled - that

@therobotjames outperformance comes from some combination of
a) multiple expansion as others buy in
b) an exogenous event (index inclusion?)
c) factor/style outperformance
d) the company hitting it out of the park, setting you up on a higher alpha slope going forward

If the answer is d), you

@therobotjames obviously stick with your name. If it is any of the others, then clearly a bunch of your recent "alpha" has come from oscillating factors - factors which, over a certain periodicity-name combination, revert.

If you recognize those shorter-term oscillator inputs to performance,

@therobotjames then you should also be able to recognize where they are not present (and eventually you start looking for where they COULD become present), and you identify whether near-term, that thing which WAS slightly less attractive has, on a risk-adjusted basis, become relatively more

@therobotjames attractive than the thing you are in, for nothing more than what the forward-looking prospects are for relative alpha outperformance because of 'a', 'b', and 'c'.

Why I say that suits "most" is that everyone's active share portfolio is different, but the only way one wins big

@therobotjames because of one's expertise at 'd' is by being a stock selection god or goddess, or being a very aggressive high beta high-momentum investor in a bull market.

I expect not all of us are either of the latter all of the time, so if we take the short step to actually believe that,

@therobotjames harvesting shorter-term input oscillation alpha should be more of a thing.

I note the higher the active share and the lower the momentum contribution to factor performance the more important that harvesting is.

@therobotjames Of course, if the model is Go With The Mo (nothing wrong with that), then the one tailors one's relative turnover (i.e. name replacement) to the periodicity of Mo With Which One Goes.

I'd also note trimming one name to add more to another which dips is "turnover" - and is a

@therobotjames decision generating higher turnover, based on timing, and assessing 'a', 'b', and 'c' as important contributors. So even in a Go With The Mo portfolio, identification of oscillator contributors to alpha and optimising periodicity and forward relative alpha is key.

@therobotjames Stupid example:
Stock A has two share classes (A and A2). The Discount of A2 to A oscillates over time between 60% and 30%.

You like Stock A, and Stock A has outperformed A2 for the last two years while you have held it.

You think Stock A has another two years before its

@therobotjames alpha vs peer/comp basket starts to wane because of difficulty of marginal margin growth, or relative product maturity, or whatever. Your expected horizon is 2yrs-ish, extendible.

The A2/A oscillation cycle, peak to peak, or trough to trough, is 2-6yrs historically and the

@therobotjames peak to current trough is 2yrs (it was a good choice to buy A vs A2 when A was at a 30% discount 2yrs ago so you are feeling good about yourself).

A2 is now at a 55% discount to A.

What do you do? How do you think about this?
1) do you care about voting rights?

@therobotjames 2) does company care about ROE?
3) what happens historically to A2/A when OPM falters?

If you can live without full voting rights (assuming A2 is a non-fungible class/pref), you might now look at A2.

Why? Oscillator factors
1) if 55% discount moves to 45% discount AT ANY TIME

@therobotjames in the next two years, you got oscillator alpha on top of your magnificent choice of owning Stock A.
2) if the A2 discount narrows when A falls because most active holders own (and then sell) A, you suffer less - at that point owning A2 vs A is something of a hedge.

@therobotjames 3) if mgmt cares about ROE and EPS smoothing/manipulation over the years, they often change capital allocation policy when R growth starts to get soft. At that point, it makes sense to buy back A2 because it is a much cheaper form of E (and no, the ROE calculation doesn't care).

@therobotjames All told, this is a matter of assessing likelihood of path and incentives over your horizon given circumstances.

This is why Prosus at 20% discount to basket just after listing was a relative sell and why others currently at near 70% are a buy.

Nobody knows whether that 70%

@therobotjames discount name will do the right thing and offer to buy back all its discounted class at a 30% discount next year (you say "gee, that's not fair" - that's 30%", but I say "I bought at 30cts on the dollar, hoping to get out at 50 and now I have full no-friction exit at 70? Done!"

@therobotjames Going from 55 to 45 as part of the oscillation experience is 22% relative outperformance.

Going from 70% to 50% is 66% outperformance.

Which is why watching the Softbank discount to NAV is key. Understanding WHERE mgmt cares (or IF they care) is key. And if you can effect a

@therobotjames change in incentive (money, pride, whatever), that's all to the good.

@therobotjames Apols this is all PM 101 to those reading this far down. My point really is that if you are active, you have to be thinking about a baseline alpha capture of far over your cost of commissions, or else just entrust it to someone else and get another hobby. And if you are, then

@therobotjames every decision you make is thinking about that alpha, which at its core is a decision about relative performance over time. If you are skilled at those decisions (or think you are), exercise that skill.

Any skill you have will far outstrip the comms you pay so any *incremental*

@therobotjames decision will also far outstrip the comms you pay. Alpha does not compound serially forever based on single stock picks unless you look at rolling 10yr periods on FANGs, and there, Mo Begets Mo - fundamentally and in markets - until one day it does not.