Fixed in­come at­tri­bu­tion in the in­vest­ment process

Finweek English Edition - - Property Compass - Adrian van Pal­lan­der Corona­tion Fund Man­agers

PORT­FO­LIO PER­FOR­MANCE mea­sure­ment tells us how a fund has done his­tor­i­cally and what the level of per­for­mance rel­a­tive to the bench­mark has been. It’s the qual­ity con­trol el­e­ment of the in­vest­ment process and pro­vides clients with the nec­es­sary in­for­ma­tion to ac­cu­rately as­sess the re­sults of that process.

How­ever, per­for­mance mea­sure­ment doesn’t tell you how the re­turns were gen­er­ated. Out­per­for­mance may have been the for­tu­itous com­bi­na­tion of in­vest­ments that didn’t form part of a rig­or­ously de­fined in­vest­ment process.

If a man­ager’s per­for­mance were ran­dom (ie, luck) we wouldn’t ex­pect that per­for­mance to be repli­cated in sub­se­quent years.

Con­se­quently, you’d want to ask if there was a way to de­com­pose a port­fo­lio’s per­for­mance in or­der to iso­late the sources of re­turn. In par­tic­u­lar, is there a way to an­a­lyse the re­turns of a fixed in­come port­fo­lio in or­der to show the man­ager is los­ing money from credit spread changes while mak­ing money from his du­ra­tion po­si­tion­ing? There is ‒ and it’s called at­tri­bu­tion anal­y­sis.

Per­for­mance at­tri­bu­tion anal­y­sis is the ex­er­cise of de­com­pos­ing a port­fo­lio’s per­for­mance to de­ter­mine how the money man­ager has achieved the cal­cu­lated re­turns for a given as­set class (for ex­am­ple, bonds) or across as­set classes.

Thanks to Gary Brin­son and many oth­ers, peo­ple gen­er­ally agree on how to cal­cu­late eq­uity per­for­mance at­tri­bu­tion and are comfortable with its in­ter­pre­ta­tion. How­ever, the com­mon de­bate is whether clas­sic eq­uity at­tri­bu­tion mod­els can be adapted for the per­for­mance of fixed in­come port­fo­lios?

That ar­gu­ment is in light of the very dif­fer­ent in­vest­ment pro­cesses fol­lowed by fixed in­come and eq­uity teams.

An­other area of con­cern is the per­for­mance spread be­tween a bond port­fo­lio and its bench­mark, which is gen­er­ally smaller than that of an eq­uity port­fo­lio and thus a higher level of pre­ci­sion is re­quired for a fixed in­come at­tri­bu­tion model.

But there will never be a “one size fits all” ap­proach to fixed in­come per­for­mance at­tri­bu­tion. Ideally, the fac­tors on which a per­for­mance at­tri­bu­tion model is based should be con­gru­ent with the in­vest­ment process the port­fo­lio man­ager pro­fesses to use.

Thus, in­stead of looking at per­for­mance in terms of ex­ter­nally im­posed sec­tors, per­for­mance at­tri­bu­tion mod­els ‒ based on the dif­fer­ent in­vest­ment pro­cesses ‒ al­low you to look at the re­turns gen­er­ated by each type of in­vest­ment de­ci­sion made by the man­ager.

One ex­am­ple of such a cus­tomised fixed in­come per­for­mance at­tri­bu­tion ap­proach is the ap­pli­ca­tion of prin­ci­ple com­po­nent anal­y­sis (PCA) on the re­turns of vanilla fixed coupon bonds and cash. PCA is a sta­tis­ti­cal tech­nique that ex­tracts the most com­mon types of yield curve shifts over a pe­riod.

The three most com­mon types of yield curve shifts are level shifts (that cap­ture the du­ra­tion view of the port­fo­lio), slope changes and cur­va­ture changes.

Sub­se­quent cal­cu­la­tions into the re­turns on the fixed coupon por­tion of the port­fo­lio are bro­ken down in re­turns due to coupon pay­ments and pull-to-par, as well as the yield pick-up due to hold­ing cor­po­rate bonds and credit spread changes. All other in­stru­ments are priced sep­a­rately and don’t con­trib­ute to the fac­tors re­sul­tant from the PCA. Their re­turns are then ap­pended to the re­turn break­down.

The graph il­lus­trates an ex­am­ple of such an at­tri­bu­tion on an ac­tive bond fund. That’s an at­tri­bu­tion of the out­per­for­mance rel­a­tive to the all-bond in­dex, which was 1,05% over the pe­riod.

If we look at the graph we see the bulk of the out­per­for­mance came from level shifts in the yield curve. That comes mainly from the mod­i­fied du­ra­tion po­si­tion­ing of the fund rel­a­tive to the bench­mark. Mod­i­fied du­ra­tion is the mea­sure of change in the port­fo­lio value due to changes in the un­der­ly­ing in­ter­est rate.

We also see changes in the shape of the curve added an­other 11 ba­sis points to out­per­for­mance, whereas the big neg­a­tive im­pact came from credit spread changes.

An in­vestor can look at that anal­y­sis to de­ter­mine whether the port­fo­lio’s sources of re­turn match the in­vest­ment process the port­fo­lio man­ager claims he’s us­ing or if they come from other sources not cov­ered in the in­vest­ment process.

If the for­mer is true it in­creases the con­fi­dence the in­vestor can have on the qual­ity and po­ten­tial repli­ca­tion of the al­pha de­liv­ered by the man­ager.

Newspapers in English

Newspapers from South Africa

© PressReader. All rights reserved.