For multiple periods. I just read the whole Study Session - I couldn’t grasp too much about multiple periods. Does anybody have a good summary/explanation for this part? I know it’s a lot to ask.
I’d remember the formula to calculate the 2 periods return. (portfolio return period 1 * Benchmark return period 2) + (portfolio return period 2 * portfolio return period 1)
^ But it’s “active return period 1”, not ‘portfolio return’. Two Period active return = (Active Ret 1)*(Bench Period 2) + (Active Ret 2)*(Portfolio Ret 1)
yeah, mcleod is right. had this wrong on a recent exam question somewhere as i recall and got stumped. k-rock
is this in the LOS? Don’t see it anywhere in my guide 
Also, 1) You can’t compound active returns for each period: 2 Period Active Ret <> (A1)*(A2) 2) You can’t add active returns for each period: 2 Period Active Ret <> (A1) + (A2) 3) You can’t assume it’s the same percentage of whatever… You CAN also subtract the multi-period benchmark return from the multi-period portfolio return 2 Period Active Return = (1+Portfolio Ret1)*(1+Portfolio Ret2) - (1+Bench1)*(1+Bench2) - 1
you guys are totally right, thanks for pointing out
Thanks McLeod I think 3 was that you can’t assume the same percentage of active return from a period to the other
Probably the most difficult topic. Does anyone have notes or something on this?
inbead Wrote: ------------------------------------------------------- > Probably the most difficult topic. Does anyone > have notes or something on this? Are you looking for “notes or something” on multi-period attribution, or performance evaluation in general? Let me know what you’re having issues with, and I’ll post some stuff.
The two most difficult concepts to keep straight are 1- multi period attribution and 2 - Global performance evaluation. It is hard to memorize all the formula and keep them straight during the exam. Thanks in advance darkstar.
http://www.analystforum.com/phorums/read.php?13,949683,949683#msg-949683 http://www.analystforum.com/phorums/read.php?13,981752,985458#msg-985458 Look for my posts from last year in these threads. That should help clear up global performance evaluation. I think it’ll just be a case of memorizing it. I’ll see if I can come up with a good example that might make the formulas make more sense. I’ll work on something for multi-period attribution. Stay tuned!
OK, so first of all, we need to understand that: 1. you can’t add active returns across periods 2. you can’t compound active returns across periods 3. you can’t assume that the proportion of active returns stays constant across periods Moving on…we now have our formula: ------------------------------------------ Two pd active rtn = [(pd 1 active rtn) * (1 + pd 2 benchmark rtn)] + [(pd 2 active rtn) * (1 + pd 1 portfolio rtn)] ------------------------------------------ [(pd 1 active rtn) * (1 + pd 2 benchmark rtn)] This term is saying that the active return in the first period must be compounded at the benchmark return in the second period. Assume you pursued an active strategy in period 1, and then moved to a passive strategy in period 2. For example, say you start with $100, and invest it in the market during period 1, and you make 10%, while the market is up 1% - so your active return is 9%. Then, you change to a passive strategy in period 2, where you start with $110. The market goes up 5%, so you’re taking your 9% active return in the first period, and compounding it forward at the market return in the second period. ------------------------------------------ [(pd 2 active rtn) * (1 + pd 1 portfolio rtn)] This term is saying that your active return in the second period must be compounded at the portfolio return from the first period. The portfolio return from the first period combines the impact of your benchmark return and your active return. Continuing with our earlier example but choose to pursue an active strategy in period two. Say the market goes up 4%, and your portfolio goes up 6%, so your second period active return is 2%. So now we need to go back and look at our portfolio return in the first period, which was 10% - so then we compound the second period active return forward at the first period portfolio return (since the portfolio return from the first period is our starting point at the beginning of the second period). Now we can reverse the active/passive thing and you can start to see how the two formulas connect. Assume that we pursued a passive strategy in period 1, and an active strategy in period 2. You’d have no active return in period 1, so the first term would be zero, and the second term would become: [(pd 2 active rtn) * (1 + pd 1 benchmark rtn)] ------------------------------------------ Does that help?
This might help make sense of the formulas. allocation = (portfolio sector weight - benchmark sector weight) * (benchmark sector return - benchmark return) The allocation effect is the difference in sector weight between the portfolio and the benchmark multiplied by the “extra” return (the amount by which each sector outperformed the overall benchmark) for each sector in the benchmark. Essentially, you’re trying to find the amount of active return you gained by overweighting or underweighting a given sector, and assumes that the actual securities that you hold within each sector are equivalent to the benchmark. selection = (benchmark sector weight) * (portfolio sector return - benchmark sector return) The selection effect is the benchmark sector weight multiplied by the difference in the sector returns for the portfolio and the benchmark. Essentially, you’re trying to figure out the amount of active returns you gained by picking different securities than the benchmark. This term assumes that the sector weight in the portfolio is equivalent to the benchmark, but the actual security weightings within each sector are equal. interaction = (portfolio sector weight - benchmark sector weight) * (portfolio sector return - benchmark sector return) The interaction effect measures the combined effect of your allocation (sector choice) and selection (stock picking) decisions. Please let me know if you can come up with an intuitive explanation for this term. Questions?
Thanks much for above, I checked out your quotes from last year as well. Here are the issues I am having: 1. Remembering when to apply which formula (knowing global vs micro attribution and decomposition vs attribution) 2. Remembering the difference between Currency Contribution and Currency Effect. Again, when to use which formula? 3. Schweser talks about Decomposition and attribution differences. I am assuming decomposition is when there is no benchmark and attribution is when there is benchmark. 4. 2007 Question, the formula used are pretty different for Attribution and Security Selection. Is that something we shouldn’t worry about as it was based on prior curriculum? 5. One last thing is while there is a formula for Pure Sector Allocation in SS46 (Micro), there is no similar formula in SS47 (Global)? Why is that? Similarly, there is no formula for Market allocation in SS46.
Attribution means that you’re finding the sources of ACTIVE return. Decomposition (or, as I would call it, contribution) means that you’re finding the sources of TOTAL return. 1. If the question specifies a global benchmark, I would go with global. If not, assume micro. If the benchmark information is given, perform attribution. If not, perform contribution (decomposition). I don’t think they’re going to make it too confusing on the exam. 2. Currency effect refers to the return from currency movements in the PORTFOLIO. Currency contribution (meaning contribution to ACTIVE return) means the effects of currency decisions in the portfolio versus the benchmark (attribution). 3. See #1 4. See page 211 in the CFA material, or page 137 in the Schweser material 5. They change the terminology between reading 46 and 47, and I think this is what has introduced a majority of the confusion with this section. Forget all their jargon for a second (really, it WILL help, I promise) and remember that active return = allocation + selection + interaction. Pure Sector Allocation from reading 46 is the same as Market Allocation from reading 47. Make sense?
Darkstar, this is good thanks. I can finally say, that I have got a good understanding of the Perf Eval topic and this is my 2nd time attempting the L3 exam. Now on a completely different topic. Do you have any strategy to memorize, remember the various option strategies? You know, profit, loss, etc. on Butterfly Spread, Straddle, etc.?
Good stuff, glad you liked it. Feel free to post any additional questions on these readings if you have any. No, I don’t. I was kinda hoping someone else did!
I would suggest only knowing HOW to construct any of these strategies… from there it is only math: 1- Value at expiration is the sum of the payoffs resulting from the options. ie, if you consider a bull spread strategy, you know it’s a long call and a short call, so the value at expiration is simply the sum of the payoffs: max(0, ST-X1) - max(0, ST-X2). 2- Profit is Value at expiration derived above minus the costs to enter the positon. The costs are always the premiums paid on the long options minus the premiums received on the short options. So for a bull spread, profit = max(0, ST-X1) - max(0, ST-X2) - c1 + c2. Just note that in the case of a collar, you have also the cost of the stock since it s a strategy that combines a stock, a long put and a short call. 3-The maximum loss is ALWAYS the net costs of the premiums on the options. In a butterfly spread for ex, where you combine one long call X1, 2 short calls X2, a one long call X3, your maximum loss is the net cost: c1 - 2c2 + c3. (except for the collar) 4-The maximum gain whenever you have spread strategies, ie strategies that combine options of the same nature (only calls or only puts), is always X2- X1, the difference between the strike prices on the options, even in the case of the butterfly where you have also X3. In the case where you don’t combine options of the same nature, like protective puts and covered calls, logic or graphs are enough to determine the max loss. 5-As for the breakeven price, although it might take a few seconds more to derive, you can take the expression of profit you derived above IF IT CONTAINS THE TERM St and set it equal to zero and solve for St. In the case of butterfly spreads for ex, you have 3 expressions for the profit according to the price of St at expiration (St
darkstar, I got following learning outcomes after my reviewing of Schweser 2009 V5 P241~244 (Example : Return decomposition) and I think they are important. Would you please check if following are correct (if no cash flow yield) : 1) If no int’l (world) benchmark (index) is provided R.p.d = R.b.d [= Sigma (Wi,b x Ri,b.d)] + SS + MA + CAC = Sigma (Wi,p x Ri,p.d)R.p.d 2) If int’l (world) benchmark (index) is provided R.p.d = IB Return + SS + MA + CAC = ÓWi,p x Ri,p.d [IB Return : (IBT/ IB0) – 1] R.p.d : domestic return on the portfolio R.b.d : domestic return on the benchmark SS : Security Selection contribution MA : Market Allocation contribution CAC : Currency Alocation contribution IB : Int’l (world) benchmark (index) IBT : Int’l (world) benchmark (index) at the terminal IB0 : Int’l (world) benchmark (index) at the inception 2. On the other hand, I am confused by following terminologies used in CFAI text & Schweser. Please confirm following. it’s better for us to clarify to avoid confusion. Capital gain = Return (in local currency or home currency, without cash flow) ? Total capital gain in (each) currency = Total return in (each) currency(without cash flow) ? Currency component = Currency contribution = Currency effect ? Market return = Benchmark’ return (for indivisual security and/or portfolio) ? 3. Old exam 2007 AM Q9 Shall we assume that the weights of local index (benchmark) are all zero ? Thank you for your help in advance !