Systematic Cash Management for Futures Traders — Part I

Margin Mechanics

Let’s start with the basics: the mar­gin­ing sys­tem. If you’re read­ing this arti­cle, you prob­a­bly already have a work­ing under­stand­ing of mar­gin mechan­ics, but let’s do a quick recap. First of all, there is an impor­tant dif­fer­ence between buy­ing stocks and trad­ing futures on mar­gin. When buy­ing stocks, you typ­i­cal­ly pay only half of the price your­self and the oth­er half is financed by a loan from your bro­ker if you run out of cash. When buy­ing futures, there is no loan involved: The bro­ker sim­ply sets aside a pre­de­fined ini­tial mar­gin from your account.

Sup­pose your bro­ker account is denom­i­nat­ed in US dol­lars (your base cur­ren­cy) and you are bull­ish on the Ger­man econ­o­my, so you decide to buy a DAX mini future. Behind the scenes, in a win­dow­less room down­stairs at your bro­ker, the fol­low­ing hap­pens: The last time you checked, your account val­ue was $100,000 and the DAX trad­ed at 23,000 index points. The con­tract is denom­i­nat­ed in euros and has a mul­ti­pli­er of 5. One con­tract there­fore has a cur­rent notion­al val­ue of €115,000. Even if your dol­lars are not worth as much as they used to be, you can afford one DAX mini. You can even afford a lot more because futures are lever­aged instru­ments that trade on mar­gin.

The bro­ker requires about €10,000 of your account equi­ty as col­lat­er­al to open the posi­tion. This is the ini­tial mar­gin required to buy one DAX mini and is usu­al­ly deter­mined by the exchange. The bro­ker can also increase this mar­gin if deemed nec­es­sary. Once the posi­tion is open, you only need about €9,000 as col­lat­er­al to sus­tain the posi­tion. This is called main­te­nance mar­gin. The dif­fer­ence between the ini­tial and the main­te­nance mar­gin can be seen as a buffer for the price of the DAX mini to move around a lit­tle bit before the bro­ker has to call you up and require addi­tion­al funds to main­tain the posi­tion.

But wait a sec­ond: You do not have a sin­gle euro in your account. And yet the trade goes through. No prob­lem! Post­ing the ini­tial mar­gin does not involve any cash move­ment or cur­ren­cy con­ver­sion in your account. The friend­ly peo­ple down­stairs sim­ply set aside a dol­lar equiv­a­lent from your account as col­lat­er­al to cov­er the ini­tial mar­gin. This pro­ce­dure only affects how many addi­tion­al posi­tions you can open. For­eign cur­ren­cy cash flows occur only lat­er on, when the future con­tract changes in price and dai­ly set­tle­ment occurs. Just remem­ber: Mar­gin is not mon­ey that moves—it’s mon­ey that gets locked.

Interest Strategy

Now that the mar­gin mechan­ics are under­stood, let’s dive into inter­est cal­cu­la­tion at Inter­ac­tive Bro­kers. Inter­est is easy to ignore—until you real­ize it has been drain­ing your account for months. It accrues dai­ly. Qui­et­ly. And then shows up at the end of the month when you near­ly for­got about it. Every cash bal­ance per cur­ren­cy is tied to the respec­tive overnight bench­mark rate—but Inter­ac­tive Bro­kers does not pass that rate through ful­ly. There are some caveats:

First, they inter­nal­ly have two seg­ments: A secu­ri­ty and a com­mod­i­ty seg­ment. Futures mar­gin sits in the com­modi­ties segment—and does not earn inter­est. But you also don’t have to pay inter­est if, say, your USD account opens a DAX posi­tion with­out hold­ing any EUR cash. Inter­ac­tive Bro­kers uses your dol­lars as col­lat­er­al for the mar­gin. They effec­tive­ly pool all cus­tomer posi­tions for mar­gin cal­cu­la­tions at the clear­ing house.

Earn­ing inter­est on non-base cur­ren­cy bal­ances there­fore is quite dif­fi­cult: You would have to have more cash in that cur­ren­cy than the main­te­nance mar­gin and also hit the cash bal­ance tier (above 10,000 USD val­ue in the respec­tive cur­ren­cy). That is some­thing you usu­al­ly should not do, because it cre­ates too much FX risk.

Sec­ond, they do not give you the bench­mark rate on cash posi­tions. They pay you whole­sale minus a hair­cut and charge you retail plus a markup while the park­ing meter is tick­ing all night. For pos­i­tive bal­ances, the hair­cut is between 0.5% for US dol­lars and 4% for Mex­i­can pesos for the low­est tier¹ to the respec­tive bench­mark rate. And because the peo­ple down­stairs want to paint the walls from time to time, they also charge you an add-on for neg­a­tive cash bal­ances (between 1.5% for US dol­lars and 3% for Mex­i­can pesos for the low­est tier²). In oth­er words: Earn­ing inter­est with Inter­ac­tive Bro­kers on cur­ren­cies oth­er than your base cur­ren­cy is there­fore often sub­op­ti­mal, while neg­a­tive cash bal­ances are expen­sive to main­tain. That’s an impor­tant point to keep in mind when design­ing the strat­e­gy lat­er on.

From a futures trader’s per­spec­tive, inter­est at the bro­ker is a game with two sim­ple rules:

Let’s say you are pret­ty pru­dent about avoid­ing neg­a­tive cash bal­ances and do not exces­sive­ly hoard non-base cur­ren­cies in your account. As a futures trad­er you typ­i­cal­ly will sit on about 70% of your cash, because respon­si­ble trad­ing at a risk tar­get of e.g. 25% will usu­al­ly only use between 20% and 30% of your cap­i­tal as mar­gin. There­fore, you could invest the 70% ”free” cash oth­er­wise.

If you do not have a spe­cial strat­e­gy for that and want to stick to liq­uid instru­ments, you could park your mon­ey in a mon­ey mar­ket ETF track­ing the bench­mark of your base cur­ren­cy (like SOFR for USD). This way you do have a risk expo­sure sim­i­lar to before (no FX risk and dai­ly inter­est rate risk). Of course, ETFs are not risk free: They add costs in the form of the Total Expense Ratio (TER)—about 0.1% per year. They also work well. Until the moment you actu­al­ly need the mon­ey. And that moment tends to coin­cide with mar­ket stress, when liq­uid­i­ty is sud­den­ly not as reli­able as adver­tised. And times of mar­ket stress could be exact­ly the times when you need more mar­gin because your trad­ing strat­e­gy tanks. In such times, bro­kers also tend to increase mar­gin require­ments on top of that. So, keep­ing a cash buffer in addi­tion to your ETF seems pru­dent.

The liq­uid­i­ty dev­il last struck the mon­ey mar­ket in March 2020, when COVID hit. At the time, nei­ther SOFR nor €STR exist­ed in their cur­rent form, but mon­ey mar­ket ETFs track­ing T‑bills or EONIA, the pre­de­ces­sor of €STR, encoun­tered dis­lo­ca­tions. Going down the ETF route and find­ing a good prod­uct for your use case requires some research. Avoid opaque swap struc­tures, high coun­ter­par­ty risk and stick with the most liq­uid ETFs—and read the fine print. Also remem­ber that let­ting your cash sit at your bro­ker is also not risk-free—even if they hold your mon­ey at sev­er­al banks.

What can you gain by going down this path? Let’s say you have a decent account size for diver­si­fied futures trad­ing of $500,000 and your mar­gin uses 30% of that. You want to also employ a cap­i­tal buffer of 20% so you can invest half of your account size in an overnight index ETF. You pick one at a TER of 0.1%.

Com­pare this with leav­ing the same cash bal­ance at Inter­ac­tive Bro­kers: You can cal­cu­late your aver­age inter­est rate at their site. Leav­ing 50% of the account as cash at Inter­ac­tive Bro­kers results in a hair­cut of 0.63% com­pared to SOFR. The ETF only has a TER of 0.1% and some trans­ac­tion costs and a spread. Your improve­ment against Inter­ac­tive Bro­kers will be about 0.50% per year.

Let’s put that into per­spec­tive to your trad­ing strat­e­gy. If you run a decent strat­e­gy with a Sharpe ratio of 1.0 at a risk tar­get of 25%, your year­ly expect­ed return turns out to be also 25%. The ETF strat­e­gy cor­re­sponds to a Sharpe ratio increase of 0.02. The improve­ment is small, but not neg­li­gi­ble as it also scales with your account size. I have put far more work into parts of my trad­ing sys­tem for gains of sim­i­lar size. In absolute terms: 0.5% of $250,000 equals $1,250 per year. That’s a non-neg­li­gi­ble part of the year­ly trans­ac­tion costs and seems worth­while.

So far, we assumed every­thing works smooth­ly. But what hap­pens in stress? To close this chap­ter, there is anoth­er very impor­tant aspect about cash man­age­ment in com­bi­na­tion with ETFs: Mar­gin calls. A mar­gin call occurs if your NLV is small­er than the sum of main­te­nance mar­gin in your account. At Inter­ac­tive Bro­kers this will trig­ger auto­mat­ic fire sales of some of your posi­tions to bring the mar­gin require­ments back in line with your NLV. Even if this hope­ful­ly nev­er occurs, your sys­tem­at­ic trad­ing strat­e­gy has to be able to deal with this.

In nor­mal times the safe­ty buffer is enough to absorb any P&L shocks but it also can go to zero. The trad­ing sys­tem needs an alert mech­a­nism that liq­ui­dates posi­tions auto­mat­i­cal­ly when the thresh­old is hit. I sug­gest that you size down your ETF posi­tion the moment your mar­gin require­ments hit 95% of your NLV. Sell as much as you need to bring your mar­gin back to 80% of your NLV. If the ETF posi­tion has been ful­ly liq­ui­dat­ed, you start liq­ui­dat­ing your futures posi­tions.

There is a small edge in clos­ing the biggest dai­ly win­ners first. That’s because a lot of instru­ments exhib­it slight mean rever­sion (neg­a­tive auto­cor­re­la­tion) in the short term. This behav­ior seems to per­sist since the year 2000³. It means that today’s win­ners tend to be tomorrow’s losers (see fig­ure 1). This effect comes in handy when you are forced to close a position—if it made mon­ey today, odds are a lit­tle bit high­er that it will lose mon­ey tomor­row.

We now have all ingre­di­ents for a sys­tem­at­ic rule set: mar­gin uti­liza­tion, ETF expo­sure, cash buffers, and a liq­ui­da­tion pri­or­i­ty. The next step is to turn these con­cepts into two sim­ple algo­rithms.

while (margin > 0.95) {
  if (n > 0)
    SELL k <= n ETF shares
  if (n == 0)
    CLOSE future with highest PnL
} until margin < 0.8

addedMargin = (200 * P) / NLV
currentETFMargin = n * P / NLV
if (margin + addedMargin  < 0.8)
  if (currentETFMargin < 0.5)
    BUY 200 ETF shares

ETF Selection

One prac­ti­cal ques­tion is still unan­swered: Which ETF should you use to invest your spare cash? The answer is dri­ven by the ETFs avail­able for trad­ing at Inter­ac­tive Bro­kers, the exchanges they trade on with their respec­tive com­mis­sion struc­ture and the typ­i­cal bid/ask spreads. The exam­ple below is EUR-focused because my base cur­ren­cy and reg­u­la­to­ry set­ting are Euro­pean; the same selec­tion log­ic can be applied to USD, CHF, GBP, or any oth­er base cur­ren­cy.

Let’s dive into the com­par­i­son. Inter­ac­tive Bro­kers offers the fol­low­ing ETFs (see Table 7). I fil­tered for accu­mu­lat­ing and UCITS com­pli­ant ETFs (they rein­vest the returns). UCITS is the Euro­pean reg­u­la­tion for exchange-trad­ed funds. All of these ETFs use swap-based repli­ca­tion through agree­ments between the fund and one or more banks. One advan­tage of UCITS is that coun­ter­par­ty expo­sure from over-the-counter (OTC) deriv­a­tives is capped: gen­er­al­ly 10% of fund assets for qual­i­fy­ing cred­it insti­tu­tions and 5% for oth­er coun­ter­par­ties. This con­strains, but does not elim­i­nate, coun­ter­par­ty risk.

This list can be nar­rowed down fur­ther: Both the BNP Paribas ETF and the Invesco ETF are rel­a­tive­ly small, so I exclude them from con­sid­er­a­tion: I do not intend to be the largest investor in any ETF!

The remain­ing ETFs use unfund­ed swap repli­ca­tion. They invest share­hold­er cap­i­tal in a sub­sti­tute or col­lat­er­al bas­ket and con­tract a swap agree­ment with a swap provider. The ETF exchanges the total return of the col­lat­er­al port­fo­lio for the €STR return (in case of the Xtrack­ers even €STR plus 8.5 basis points).

This exhibits two main risks: Coun­ter­par­ty risk—the risk that the swap coun­ter­par­ty defaults or oth­er­wise fails to per­form under the swap. And col­lat­er­al risk: This is the risk that the col­lat­er­al falls in val­ue. Col­lat­er­al risk becomes rel­e­vant pri­mar­i­ly in the event of a coun­ter­par­ty default, although dai­ly mar­gin­ing sig­nif­i­cant­ly reduces this risk. The qual­i­ty of the col­lat­er­al is espe­cial­ly rel­e­vant in sys­temic crises. Swap coun­ter­par­ties are usu­al­ly banks which only default in finan­cial crises and in such a case it is safer to have gov­ern­ment bonds as col­lat­er­al than US tech com­pa­nies.

From a risk per­spec­tive (and we want to have as lit­tle risk as pos­si­ble in our cash invest­ment) it makes sense to invest in an ETF that uses bonds as col­lat­er­al. The Xtrack­ers offers the most diver­si­fied and con­ser­v­a­tive col­lat­er­al struc­ture and the largest fund size, which can be advan­ta­geous from a risk and liq­uid­i­ty per­spec­tive. Its return is also slight­ly high­er with €STR plus 8.5 basis points. The Xtrack­ers also trades on Xetra in Frank­furt which exhibits the low­est com­mis­sions of the rel­e­vant exchanges and is trad­able at a low bid/ask spread. I guess we found a hot can­di­date!

Part II will intro­duce a strat­e­gy for con­vert­ing non-base cash bal­ances more effi­cient­ly. It will also teach you, how to aggre­gate trad­ing sig­nals and your cash posi­tions for slight­ly bet­ter returns and less trad­ing.

Appendix A: Definitions and Methodology

Appendix B: Research and Publication Conventions