Transaction costs - subscribers: how much are you investing?

Models
1

I am interested in doing some “transaction cost analysis”, as it is clear to me a lot of subscribers are concerned about transaction costs, and rightly so.

Going right back to the implementation of Variable Slippage, I argued against making transaction costs mandatory for R2G portfolios, (beyond the slippage factor), as everyone faces a different set of transaction costs.

There is a basic problem simulating fixed rate commission - if you backtest from 1999, by current day (2014) compound growth means the portfolio value is massive compared to the original value, and the commission has a negligible impact on returns. To fix this, I have converted the fixed rate commission into a percentage for the purposes of simulation.

E.g. You have $50,000 to invest, portfolio has 20 positions, and the commission rate is $10/trade.

Position size = $50,000 / 20 = $2,500
Commission as a percentage = $10 / $2,500 = 0.4%

Now a simulation can be run with a commission set to 0.4% of the trade value, this way the relative impact does not decline as the portfolio grows.

The “problem” is that the impact of fixed-rate commission is a function of both the commission rate and the portfolio size. There is no “one size fits all” answer.

I have decided it might be useful to offer an analysis showing the annualised return for a range of portfolio sizes, and commission rates with the information presented in a table.

I have done this for a couple of my model portfolios that you can see here and here . A combination of low portfolio size, high commission rate and high turnover may completely wipe out the potential gains, so I think this is useful information.

I have picked a range of portfolio sizes, $10k, $20k … $100k

I have picked a range of fixed commission rates that are being charged by popular brokers: $1.00, $4.75, $6.95, $7.99, $8.95, $9.99.

I would very much appreciate some feedback on the usefulness of this approach. In particular, right now I am guessing at how much the average R2G subscriber is investing per model. If the amount is >$100k, it probably doesn’t matter as the impact of commission is fairly small. However, I could also provide analysis for impact sub-$10k. Would that be helpful for anyone?

Also, is there any other commission rate that would like to be examined?

2

Is a fixed commission rate not simply a fixed commission rate?

So if Scottrade (for example) charges $7 per trade is that not $7 per trade regardless of trade size?

3

I use Scottrade and pay $7 per trade. I’m subscribed to 3 R2G’s, 5 stocks each. I put $20,000 into one of them. The other two I havn’t started trading yet.

Commission as a percentage = $14 / $4,000 = 0.35%. This seems very low, but if the portfolio trades flat or down, this eats away the funds. If the portfolio goes up significantly, then the commisions is nothing.

I look at Average Return per trade and require it to be > 4%. This number is a bit arbitrary but it should give some margin of safety. It does eliminate some otherwise good looking R2G’s becauase they have Average Return between 3 and 4%.

I’m bounded by two limits: allocate too little funds, and fixed commisions are a concern. Allocate too much funds and the micro caps are illiquid. (Havn’t run into this second problem yet I don’t think.)

4

The question I am trying to answer is “what is the drag on portfolio returns?” In that case, percentages matter. For simplicity let us imagine a portfolio with a 100% turnover, and doubles in value each year. (after costs)

Year 1: $10k invested into 10 positions - half replaced (20 trades) - cost $140. Drag = 1.4%
Year 2: $20k - 20 trades - cost $140. Drag = 0.7%
Year 3: $40k - 20 trades - drag = 0.35%
And so on.

So by the time we hit year 15 (now) the portfolio is worth $32 million and the $140 transaction cost is just 0.0004% of the value - negligible.

If we are asking the question, “my portfolio is $10,000 and I have a $7 transaction cost”, then I don’t think running a simulation with fixed commission is going to tell us much useful. It very much depends on how the portfolio performs in the first year, if it manages to overcome the drag of commission, then the portfolio value will face less drag in subsequent years. However, if the simulation had started in year 2, if that had been a worse year, it is possible the commission drag would have been too much, and the portfolio would have drawn down and never recovered, particularly because it will face an even larger hurdle overcoming commission drag.

5

I use Vanguard and pay $2 per trade. I have four R2Gs and invest over $100k for each. I do look at average days held (and turnover) as a criteria and all are over 70 days so I don’t think commissions are a big deal for me.

6

Commissions obviously have a big effect on small accounts, but how about the impact of slippage on very large accounts (7 figures or higher)? Let’s assume that some models become so successful (just like in backtesting) that people manage to grow their accounts into the tens of millions of dollars. At that size it becomes much harder to get in and out of trades, even on many large caps. If you are doing the trades manually you have to execute multiple orders and read the level 2 carefully so that you don’t impact the price too much. And leaving a huge limit order sitting at a particular price isn’t always a great idea (The sharks WILL notice you and attack if your size is large enough).

7

This is a good question but difficult to simulate. It is one of those problems that we cannot really know in advance, and is an example of where the market feeds back on itself, and cannot be studied like an isolated system.

I guess that is where we need people to feed in their experiences, and there is the “rule of thumb” about not trading more than 5% of the volume, but that is obviously a “rule of thumb” and comes with all of the inherent limitations.

8

Oliver,

What you have described makes a great deal of sense and that is what I do.

9

Oliver has a good point. I have many times recommended that we run Sims to determine our potential percent returns for the dollar amount we want to invest now and not what we might gain in total dollars in 10 years. If we want to invest $10,000/stock in the best size stocks for that amount now, we need to know what the gains were if we had invested in those same size stocks over time. That is why using liquidity rules like AvgDailyTot(60)>xxxxxx works better than PctAvgDailyTot(20) < x for our current needs.

But if we really want to determine how investing in our current dollar amount would have performed over the past years we need to also use a % Fee instead of a fixed fee. That way the % gain per stock will more accurately reflect the fees that would have been paid if we invested a fixed amount/stock over the years.

However, I would use a fixed fee in Ports. That assumes that the Port is trading approximately the dollar amount/stock as our real trades.

Denny :sunglasses:

10

Oliver, could you draft up a standardized form we R2G subscribers could fill out and send back to you anonymously? I don’t really want to go into the details of my circumstances in a public forum, but I appreciate your efforts and am willing to contribute my data.

11

I’ve been running 6 to 8 live ports, some mine some R2G. Starting value usually $100K, sometimes only $50K if I’m not too excited about the system. But never less than $10k per holding. My trading fees are 7.95 per.

12

Oliver and others,

To decrease or eliminate the commission drag, particularly if one has a small account size, one should consider:

  1. Utilize lower turnover ETF systems at a broker with many commission free ETF: Schwab, TDAmeritrade, or Vanguard
  2. Utilize lower turnover stock systems at a less expensive broker: Scottrade, Tradeking or Interactivebrokers (Folio may also be a good fit for this space as well as the ETF space depending on the account size)
  3. Utilize a combination of 1 and 2 (an advantage to this approach is that you do not need to pay the extra fee for the live quotes at Interactivebrokers as you can access them via your second brokerage account)

After one has acquired a substantial investable amount with the above approaches they can transition to higher turnover systems without as much of a commission drag (assuming they are using a broker with a fixed cost per trade).

Note that Vanguard does charge $ 2 per stock trade but in order to get this rate one must hold >= $500,000 in vanguard funds, which works great if one has a buy and hold component, but not so well if they do not. Otherwise Vanguard charges $7 per trade but their brokerage has less features than some of the others listed above.

Scott

13

When I am deciding on subscribing to an R2G I make the following calculations.
.

  1. I will probably only subscribe for a year or 2 so the impact of commisions only matters for my initial portfolio investment and 1 or 2 year of gains. It the R2G gains a lot then the commission drag will become negligible.
  2. I estimate the drag from commissions by calculating the COMMISION % PER TRADE CYCLE= (COMMISION COST*2)/AVERAGE POSITION AMOUNT)*100.
    The AVERAGE POSITION AMOUNT is the amount at the start of subscribing to the R2G.
  3. I get the COMMISION DRAG % by calcuating the (COMMISION % PER TRADE CYCLE /AVERAGE RETURN % PER TRADE)*100.
  4. If the COMMISION DRAG % is greater than a few % then I it is a negative for the R2G.

Example: COMMISION = $10, AVERAGE POSITION= $20,000, and AVERAGE RETURN % PER TRADE = 10% then
COMMISION % PER TRADE CYCLE = (102100)/20,000 = 0.1%
COMMISION DRAG % = (0.1*100)/10 = 1%
which would be a positive for the R2G.

So the critical R2G number is the AVERAGE RETURN % PER TRADE.
A graph of the COMMISION DRAG % vs. the portfio size with multiple lines in the graph for different common commision amounts would be a good way to illustrate this for subscribers.

Jim

14

I use FolioFn, so commissions are zero. The downside is that I cannot achieve the Next Open price for executions. I believe that R2G uses the Next Average of High / Low from R2G launch, but not from simulated inception, which may be an area of inconsistency.

Regarding portfolio size, this depends on my assessment of capacity:

  1. I would like to avoid representing more than a certain percentage of average daily volume (e.g., 1%). For example, if minimum liquidity is $300,000 for a 10 stock strategy, I assume that capacity is $30,000, or ($300,000 * 1% * 10) I know that the minimum liquidity metric for R2G strategies has its flaws, as has been discussed on the forums, but this is the only data available at this time.

  2. For R2G strategies, I initiate a portfolio based on a lesser amount than my capacity calculation above, to account for some (hopefully) appreciation. Once the strategy hits my capacity limit, I will need to either subscribe to more R2G strategies, increase the holdings for my own ports, or develop new models for myself.

  3. I also assume that subscribers are investing enough to justify the monthly fee. Using the hedge fund “2 and 20” fee structure as a reasonable example, I assume that subscribers are willing to pay no more than 2% fees. For example, if the monthly fee is $50, I assume an average portfolio size of $30,000, or ($50 * 12) / 2%. I also assume a minimum portfolio size of $X for very low fee or even free R2G models.

  4. I then try to avoid strategies that are oversubscribed. After determining the average subscriber portfolio size above, I then multiply by the total number of subscribers to determine how much the R2G strategy represents of the lower liquidity stocks. For example, if minimum liquidity is $300,000 for a $50 monthly fee and 5 subscribers, I assume that R2G subscribers account for 50% of daily volume for the lowest liquidity stocks, or ((($50 * 12) / 2%) * 5) / $300,000.

15

A possible way how we handle this is that r2g models get the aumomated trading Feature with Interactive Brokers and that the model designer would be able to Charge based on a percentage of the invested capital + a fixed amount per r2g model + would be able to Limit the Overall amount of invested capital (e.g. the sum of all port sizes that invest) + a Limit of the port size per Sub.

The cost- (and Utility-!) -driver for a r2g model is the invested capital.

Lets say we look at https://www.portfolio123.com/app/r2g/summary/1141598

I would Charge 10-20 Bucks fixed per month and 0.25% of the invested capital (per year) and I would Limit the Overall invested capital in the range of 250 - 300k (I could get even higher if the the orders could be spread from monday to friday)

That would make sure that small Investors get a
great Chance to build up their Portfolio and the bigger ports do not eat up liquidity.against the smaller ports paying (relativly) much less.

I do have nothing against big ports but every port size should pay for there luquidity Needs.

for the Designers who still want the flat fee this Option could stay.

Regards

Andreas

16

Sorry I meant not 0.25 % per Month but per year!!!