position based trading strategy for leveraged etf

1 – Introduction

1Exchange-traded finances (ETFs), also called trackers in Europe, are similar to bilateral funds only their shares force out be traded at any time happening an exchange. The first ETFs have been premeditated to track an index and, as so much, have been resistless funds. Since the middle 2000's, actively managed ETFs have appeared on the market, with leveraged and inverse ETFs designed to achieve multiple exposure (positive or negative, e.g., exposures up to 2x operating room -2x the benchmark ETF) to index returns, on a each day basis. The first leveraged and inverse ETFs were launched 'tween June 2006 and June 2009 in the U.S. [1] and are become rather popular. More 150 leveraged and inverse ETFs are presently traded in the U.S., covering a thick roam of equity, sphere, international, invariable, commodity and currency markets.

2As mentioned by Charupat and Miu (2011), while the sum of leveraged funds is relatively diminutive compared to the whole ETF market, their trading volume and value are very large (for example, in Canada, during September 2009, the quantity of leveraged ETFs was about 9% of the solid ETF market while the percentages of their trading book and valuate were respectively up to 67% and 48%). Ramaswamy (2011) mention also that, "while the assets under management of leverage or inverse ETFs amount to only around $40 billion globally, which is about 3% of ETF assets, they account for nearly 20% of the turnover in ETF assets, suggesting that they are very actively listed." This is mainly due to the fact that such speculative financial products come to short-term traders with portfolio rebalancing on a every day basis. In France, a leveraged ETF which differed slightly from the US model, was introduced connected the grocery store past SGAM AI [2] in October 2005.

3Recently, there has been some controversy surrounding leveraged ETFs in the U.S. grocery store, focused mainly on the performance results delivered by these products over spread periods of time. The most frequent critique has been that leveraged ETFs practise not perform as they should operating room as they are claimed to. The Financial Industry Regulatory Authority (FINRA) published in June 2009 a regulatory notice containing the followingdannbsp;: "Piece such products Crataegus laevigata be multipurpose in some sophisticated trading strategies, they are highly complex fiscal instruments that are typically designed to achieve their stated objectives on a daily basis. Collect to the effects of compounding, their performance over thirster periods of time derriere differ significantly from their stated daily objective. Thence, inverse and leveraged ETFs that are reset daily typically are inappropriate for retail investors who program to hold over them for longer than one trading session, particularly in volatile markets." For the moment, this controversy has not crossed the Atlantic. In Europe, NYSE Euronext publishes leveraged indices [3] (which belong to the family of scheme indices) on respective national European fairness indices (CAC 40, AEX, BEL20, PSI20) which are designed to be the fundamental indices for ETFs. It gives the shadowing definition of the double leverage index number: "The leverage index number tracks the performance of a strategy which doubles exposure to an underlying index with the support of a short-run financing"on a daily basis. For example, happening the French market, happening February 10, 2010, ComStage (a secondary of Commerzbank) issued a leveraged ETF on the CAC 40 that is aimed at replicating the ETF CAC 40 Leverage index, an index launched on December 21, 2007 by Euronext.

4It is Worth noting that in France, unlike the U.S., some of the first of these leveraged funds have been managed with a slightly qualified Continuous Proportion Portfolio Insurance (CPPI) typewrite scheme [4]. Examples include SGAM ETF Leveraged CAC 40 and SGAM ETF Carry CAC 40 launched on Oct 19, 2005 away SGAM Unconventional Investment [5]. The first of these funds is not strictly speaking a leveraged ETF, since the leverage is at the most 200% but put up be less, depending on market fortune and also on the expectations of the portfolio director [6]. In what follows, we analyze these funds in Thomas More details. Avellaneda and Zhang (2010), Jarrow (2010) and Giese (2010), have recently studied these financial products mainly on a theoretical basis. Their studies only if explore the US lawsuit and none of them draws parallels with CPPI-eccentric strategies, nor considers products like that issued by SGAM AI.

5This wallpaper is methodical as follows. Incision 2 presents an introductory example which sets the background of the paper and illustrates differences between standard LETF and Leveraged CPPI strategy in a discrete-time setting. In Section 3, we analyze the value process of a leveraged ETF that corresponds to a convex constant allocation portfolio scheme. We study main of its statistical properties and examine its return as a function of the high-risk asset payof. Atomic number 3 a spin-off, we show that the stock market index price can increase piece, at the same time, the leveraged investment trust decreases. We provide the chance of such puzzling event and detail its main sensitivities to financial and management parameters. In the around-the-clock-time framework, we prove an equivalence resultant role stating that a leveraged ETF can be also be viewed As a CPPI fund with a base proportional to the portfolio value itself. In Section 4, we compare Leveraged ETFs and specific Leveraged CPPI strategies [7] in a more practical framework. First, we check a quasi-univocal expression of the Leveraged CPPI appreciate, winning chronicle of the discrete-time variations of the leveraging level. Then, we demeanor comparisons by substance of Monte Carlo simulations. Last, we use Omega and Kappa performance measures to compare both strategies by substance of Monte Carlo experiments.

2 – Prefatorial example and motivations

6There is a widely held misconception of what exactly leveraged ETFs allow: returns of 2x or 3x the annual return of the implicit market index (which corresponds to a unchangeable leveraged ETF)dannbsp;? To date, leveraged ETFs have been designed to have daily returns that are a positive or negative multiple of the daily repay of the underlying index (which corresponds in discrete-time to a daily leveraged ETF). In what follows, we recall shortly the mechanisms of inactive and daily leveraged ETFs jointly with the standard and leveraged CPPIs.

7The static leveraged portfolio strategy is a buy-and-hold strategy that aims to reproduce the leveraged performance of the stock market index over a given stop without continuous-time rebalancing (for practical purposes, day by day). This can be considered equally a bench mark for the rice beer of comparing. [8]

8Suppose that the investor trades in distinct-time. Denote by X_k the arithmetical return of the hazardous asset betwixt times t _k-1 and t_k . We have:

9

10Let us denote V^SL ₀ the value process of the static leveraged ETF at time 0 where L is the leverage coefficient of the initial portfolio valuate. This initial portfolio value is split into deuce amounts:

11

12where LV^SL ₀ is the amount (or "pic") endowed on the risky plus S and (1 ? L) V^SL ₀ is the quantity invested on the risk-free asset B. We denote by r the risk-free rate on the time period [t_k ,t _k+1]. Therefore, at any time t_k , the static leveraged portfolio return is donated by:

13

14which implies (for short horizon):

15

16In that showcase, the (arithmetic) rate of return of portfolio value is approximately equal to the rate of risky plus generate multiplied by the leverage coefficient L.

17The daily leveraged ETF is a dynamic strategy the performance of which is calculated each day according to the following equation:

18

19The Leveraged CPPI is a simplified strategy to allocate assets dynamically over meter. The Leveraged CPPI is supported on a constant proportion of the portfolio economic value invested with on the risky asset. For the criterional CPPI scheme, this constant proportion is defined on the excess assess of the portfolio with a given floor. We describe hereinafter these two methods: [9]

• The investor starts by setting a floor F_k equal to the lowest acceptable treasure of the portfolio at time t_k . It substance that the guarantee constraint is to keep the portfolio value V_k higher up the floor F_k whose dynamics is presented by:

• Then, the investor computes the buffer C_k as the excess of the portfolio esteem over the floor (C_k = V_k ? F_k ) and determines the sum of money invested in the risky plus by multiplying the soften by a preset four-fold, denoted m. The total amount invested in the risky asset is titled the photo e_k (e_k = megacycle_k ). The remaining funds (V_k ? e_k ) are invested in the riskless asset.

Consequently, the high-power development of the portfolio value is acknowledged by:

20

21where the cushion value is circumscribed past:

22

23which implies:

24

25A modified version of the standard CPPI method leads to the Leveraged CPPI (LCPPI). This method acting has a floor that is constant over time: F_k = F ₀ = ?V ₀, ?k. The parameter ? corresponds to the initial proportion of the portfolio value that is designed to comprise guaranteed. However, contrary to usual portfolio insurance, the initial exposure to the unsound asset is importantly higher than the portfolio value (200% for example), such that a leveraged portfolio is obtained. Then a maximal constipated equal to the first exposure, 200% for example, is set apart happening the risky asset photo. This bound titled the leverage is denoted by L. Therefore, the exposure is now minded by:

26

27with

28

29In what follows, we denote by LCPPI the Leveraged CPPIs corresponding to ? = 50% and away LCPPI 1 the Leveraged CPPIs corresponding to ? = 75%. If we fix the leverage L at the level 2, we get:

1. m = 4,? = 1/2. This corresponds to LCPPI fund, in the close numerical comparison.
2. m = 8,? = 3/4. This corresponds to LCPPI 1 investment trust.

Note that, for the Leverage ETF strategy, the photo is proportional to the portfolio value at whatsoever meter. The proportion L corresponds to the leverage as soon as we set up L higher than 1; for the Leveraged CPPI (LCPPI), the level is proportional to the initial portfolio assess. Thus, under condition L = m(1 ? ?), the LETF and the LCPPI are identical on the first subperiod. They differ from the second subperiod since for the LETF the amount is proportional to portfolio evaluate at time 1 (e ₁ = 55 ₁) patc for the LCPPI the amount is still partly supported the portfolio prize at time 0 (e ₁ = Min dialect [m × (V ₁ ? ?V ₀); LV ₁]).

30We illustrate all these features in the following example which considers three service line scenarii: an upward trending grocery (+5% per Clarence Day), a downward trending market (–5% per day) and a volatile market where a +5% increase is followed aside a –5% lessening [10]. While unrealistic, this data set is chosen for expositional purposes. However, note that the comments successful from this example do not depend upon these specific parametric quantity values. For the first two scenarii, the volatility of the daily comeback is equal to zero, whereas for the last scenario the daily volatility is 5% [11]. The first newspaper column contains the underlying index price, the next two columns show the leveraged (2x) ETF and the fund whose carrying out is double that of the index over the whole menstruation (static leverage). Columns (4) and (7) she a leveraged CPPI (LCPPI) portfolio which is studied to have maximum exposure to the risky asset of 200% of the total portfolio value. We also reputation the value of the exposure (columns (6) and (9)) if it had not been bounded in a higher place. The portfolios differ in the parameters used in the setting of the portfolio insurance scheme: LCPPI (resp. LCPPI 1) has a constant floor of 50% (resp. 75%) of the portfolio value at inception and a multiple of 4 (resp. 8). The initial exposure to the risky plus of these two portfolios is set apart at 200%. Then a maximal bound of 200% is localize on the risky plus pic. However, note that these funds can reach to a lesser degree 200% risky plus exposure during their lifetime. The dispute between these 2 funds is that the one with the multiple m of 8 is more than sensitive to the fluctuations of the unsafe asset (its payoff is more convex in the assess of the underlying index [12]) than the one with m = 4. Note also that the risk that this fund falls below the ball over is higher. Prorogue 1 reports data for this example. We comparison first the performance of Leveraged ETF and Static Leveraged funds. In a trending grocery store (upward or downwardly), the Leveraged ETF takes reward of the daily compounding of the leveraged exponent reappearance and exhibits a higher ternion-day restoration (33,10% in an upward securities industry and –27,10% in a downward market) than the return of the static leveraged fund (31,53% and –28,52%).

Defer 1

Effect of Return Volatility and Compounding on 2 × leveraged and LCPPI Pecuniary resource

Effect of Regaining Volatility and Compounding connected 2 × leveraged and LCPPI Funds

31Thus, in both cases and as we would expect, the leveraged ETF performs better than the static leveraged fund. Banker's bill too that, in some cases, the return unpredictability of the underlying index is nil. In a volatile market office, the leveraged ETF return may be to a lesser degree the return of the static leveraged fund, whether the underlying indicator experiences an increase (as in Table 1) or a decrease. The longer the bribe-and-maintain stop, the stronger this outcome, as we will see beneath. Volatility in financial markets has reached a very high level since the fall of 2008, which explains why investors may suffer very bad experiences with so much monetary resource. In an upward trending marketplace, both LCPPI and LCPPI 1 finances have the same carrying into action as the Daily Leveraged store due to the bound along their exposure. But, in a downwardly trending market, these two funds take advantage of their veer-pursuing sport past bit by bit cutting their exposure to the risky asset. Thus, they dampen the effect of the diminish and exhibit better performance than the Day by day Leveraged fund. LCPPI 1, with a high multiple (m = 8) than LCPPI (m = 4), amplifies this effect. In the end, LCPPI strategies behaves rather poorly in volatile markets without trends. [13]

3 – The Leveraged portfolio esteem

32In this section, first we recall that the leveraged portfolio strategy belongs to the class of the constant allocation portfolio strategies. Next, we demonstrate subordinate which assumptions a leveraged portfolio becomes a CPPI typewrite portfolio in a ceaseless-time framework.

3.1 – The value of the leveraged and pay (i.e. inverse leveraged) scheme As a staunch allocation (constant-mix) portfolio strategy

33A constant allocation (as wel named "constant mix") portfolio strategy with 2 asset classes is a dynamically rebalanced strategy that aims to maintain exposures to both asset classes proportional to the portfolio note value at some time.

3.1.1 – The financial market

34We adopt a simple continuous-time model where the stock index price [14] dynamics is given by the next stochastic processdannbsp;:

35

36which implies:

37

38where (W_t ) _t is a standard Pedesis with respect to a given filtration (F_t ) _t , S ₀, is the initial stock market index toll and ?(.) and ?(.) are respectively the drift and volatility functions of the stock market index price. These cardinal processes are fictitious to be adapted to the filtration (F_t ) _t and fulfil usual assumptions. In that framework, we can consider, for instance, stochastic volatility models, A in Cordell Hull and Edward White (1987) and Heston (1993). Note that when ?(.) and ?(.) are constant, we recover the geometric Brownian motion (GBM afterlife) as a peculiar case.

39The riskless asset price at time t is denoted by B_t :

40

41where r is the instantaneous riskless interest rate, which is assumed to be constant.

3.1.2 – Value process

42The time period considered is [0,T] and the strategies are self-funding. We consider the constant allocation portfolio strategy in which the constant proportion L of first riches V ₀ is invested in the risky plus patc the remainder proportion (1 ? L) is allocated to the unhazardous asset. Thus, the initial portfolio value is given by:

43

44where and are the first Book of Numbers of shares in the risky and riskless asset severally. The portfolio is continuously adjusted so American Samoa to maintain the exposures to each plus proportional to the portfolio value.

45Avellaneda and Zhang (2010) and Jarrow (2010) prove [15] that the value of a constant allocation (invariable-mix) self-financing portfolio scheme at any metre is disposed by:

46

47This equation can be rewritten arsenic:

48

49From Relation (17), we can see that, depending on the value of L, the portfolio value V_t is a lentiform or concave function of the gillyflower index price S_t :

50• If 0 danlt; L danlt; 1, V_t is strictly pouch-shaped.

51Usually, constant quantity-mix strategies are conceived for proportion L between cipher and one. This is why these strategies are usually referred to every bit concave strategies. This strategy performs well under relatively flatbed only volatile market conditions. Moreover, information technology capitalizes connected price reversals.

52• If L danlt; 0 or L dangt; 1, V_t or is strictly convex.

53This case can be reduced without loss of generality to the case L dangt; 1, the one that concerns us when considering leveraged funds [16]. For instance, if L is put at 2, the portfolio holdings in the ancestry index are twice the portfolio's assess. The leveraged part is financed away borrowing V_t at the riskless rate.

54Note besides that with the index price kinetics given in (11), the LETF value is always above zero. But, as shown in Appendix A, this property is no longer true for the static suit, for which continuous-time rebalancing is not allowed. Such difference betwixt the separate-sentence and nonstop-time cases is cured-known for the CPPI strategies.

3.1.3 – Whatever statistical properties

55Using Relation (16), we deduce:

56Proposition 1. – If ?(.) and ?(.) are deterministic, the leveraged ETF valuate has a Lognormal distribution:

57

58Remark 2. – In the GBM case (? and ? constant), we have:

59

60Figures 1 and 2 display the probability density function (pdf) and the additive statistical distribution social occasion (CDF) of the hark back on the leveraged ETF for different values of the leverage parameter L in the GBM case.

Figure 1

Cdf of for various L (t = 1)

Cdf of for various L (t = 1)

Figure 2

Cdf of for various L (t = 1)

Cdf of for various L (t = 1)

61As shown in Chassis 2, the high the level L, the higher the probability to start significant positive returns but also the higher the probability to get portentous negative returns. For case, for L = 2 (resp. L = 5), the probability that the return is higher than 50% is up to nigh 20% (resp. 30%) and the chance to beget a return value smaller than –50% is up to about 0.05% (resp. 30%).

62Using results about Lognormal distributions, we get:

63Proposition 3. – If ?(.) and ?(.) are settled, the first four moments of the leveraged ETF value are relinquished away:

64

65Note that, if for some s, ?_s dangt; r, both expectation and discrepancy are increasing functions of the leverage parametric quantity L. This implies that atomic number 102 such leveraged ETF dominates another with respect to the mean-variance criterion. The (relative) lopsidedness and kurtosis are as wel increasing with respect to the purchase parameter L.

66In Hold over 2, the initial cardinal moments of the comeback of these leveraged ETFs are displayed. They illustrate the stiff effect of purchase happening these statistics. The following values are secondhand for the financial market parameters: and ? = 8%, r = 3% and ? = 20%..

Table 2

Offse Four Moments of V_t/V ₀ ? 1 for versatile L (t = 1)

First Four Moments of V_t/V ₀ ? 1 for various L (t = 1)

67From Par (16), if ?(.) and ?(.) are deterministic, the expected continuous meter growth rate over the period [0,t] is equal to (experience Giese (2010a) when ?(.) and ?(.) are constant). Apart from fourth dimension, it has trio components [17]:

The variance correction terminus can even be much that the expected logreturn becomes negative. It is maximized for the rate of the purchase given by:

68

69where it reaches the respect . Relation (18) shows that the optimal leverage is stochastic and corresponds to a Sharpe-typewrite ratio. For the GBM caseful, this optimal raze is so constant. For instance, for previous business parametric quantity values, we obtain: L* = 1.25.

70It is mathematical to give other but equivalent interpretation of the prise process, supported relation (17). The growth return of the LETF over the historic period [0,t] is equal to the underlying exponent return compounded L times remittent by the concern postpaid on [0,t], exp[r (1 ? L) t], and by the effect of the ETF's excitableness on .

71As already shown by Bertrand and Prigent (2003, 2005) for the CPPI case, the portfolio value conferred away Relation (17) is inversely connate the unpredictability of the risky plus. In opposite words, the Vega of the leveraged ETF is negative. Eminence that it is the realized volatility that matters here, and not the expected (operating room implied) volatility, as is the case for choice prices.

3.1.4. A – worrying feature about LETF and risky plus returns

72In what follows, we examine how the leveraged fund value varies reported to the risky asset fluctuations, for the GBM case.

73From Relation (17), we deduce:

74

75Therefore, in continuous-time, the portfolio logreturn is equal to the risky plus logreturn increased by the leverage parameter L to which a correction term is added (this is due to the compound rates and involves the unpredictability term). The impact of this latter term has to be emphasized. Indeed, the portfolio take back is an increasing use of the hazardous asset yield but the correction term is antagonistic for values of the leverage coefficient L higher than one.

76The first implication is that the logreturn of the fund V is always smaller than L times that of the asset S. However, it is non always genuine for the comparison between the arithmetical return on the fund value V and L times that of the asset value S.

77Proposition 4. – The chance P ₁ that the arithmetical restitution connected the fund value V is less than L times that of the asset value S is disposed by:

78

79where ? denotes the cumulative distribution function (cdf) of the standard Gaussian distribution and where x ⁽¹⁾ _a,L and x ⁽²⁾ _a,L are the ii solutions of equation: , with .

80Proof. – Denote . Then, from (17), we get:

81

82Note that a danlt; 0 arsenic soon As L dangt; 1. We deliver to examine the value ?Y ? 1 danlt; L(X ? 1)]. Condition Y ? 1 danlt; L(X ? 1) is equivalent to:

83

84The function f_a,L (x) = x^Le^a ? L_x + (L ? 1) has a negligible at exp[?a/(L ? 1)] on R ⁺ with f_{a ,L} (exp[?a/(L ? 1)]) danlt; 0. Since f_a,L (0) = L ? 1 dangt; 0 and , on that point exists exactly 2 non negative numbers x ⁽¹⁾ _a,L and x ⁽²⁾ _a,L so much that:

85

86Using the equality where Z has a standard Gaussian distribution, we prove Relation (20).

87Corollary 5. – Consider the case L = 2. From Suggestion 4, we deduce that the value of the chance P ₁ that the arithmetical return on the fund prise V is fewer than L times that of the plus value S is equidistant to:

88

89Proof. – It is sufficient to determine the expressed solutions x ⁽¹⁾ _a,2 and x ⁽²⁾ _a,2 of equating: x ² e^a ? 2x + 1 = 0. We birth:

90

91with a = ?[r + ? ²]t.

92For example, for ? = 8%, ? = 20%, r = 3%, t = 0.5, L = 2, we father P ₁ ? 92%.. Figure 3 shows that the probability P ₁ is decreasing with respect to the volatility ? just is less susceptible to the expected return ? and prison term.

Figure 3

Sensitivities of P ₁ w.r.t. two parameters

Sensitivities of P ₁ w.r.t. deuce parameters

93The second and chief significance is that the threadbare forefinger can experience an step-up while at the same meter the leveraged fund decreases. This is an upshot that is in all probability thorny to take over for an case-by-case who invests on such fund.

94Proposition 6. – The chance P ₂ that the stock index increases spell the leveraged fund decreases, denoted past P ₂, is given by:

95

96Banker's bill that, past construction, we have always P ₁ ? P ₂.

97Proof. – With the synoptical notations A in previous proof, we are looking the value ?[(X dangt; 1) ? (Y danlt; 1)]. We detect that

98

99Thus, we must evaluate the probability that 1 danlt; X danlt; exp[?a/L], which is equivalent to: 0 danlt; Ln(X) danlt; ?a/L.

100Therefore, we deduct:

101

102We can gain some hunch on this phenomenon by recalling that V is resolution of the following random differential equation:

103

104Thus, even S if increases (merely not sufficiently), the shortened position happening the riskless asset B can induce a decrease in LETF value. [18]

105We can also facial expression at equation (17), which provides the LETF (compound) value at any time of the direction flow. On the unmatched hand and in the consequence that the return on the risky asset is overconfident, the first base term in the right-handed side of equation (17) tells us that the deliver on the investment trust is greater than that of the risky asset because of the leverage effect. On the different hand, the minute term is little than one. Thence, the product of these ii terms might be less than the return of the hazardous asset. This is symmetric more probably that the price S_t is ground-hugging, that the volatility, the leverage, the safe rate and the time are high. A situation where the terminal cost may be low and the volatility is high, is typically what financial markets have experienced recently.

106In what follows, we examine individual sensitivities of the probability P ₂ that the stock index increases spell the leveraged fund decreases.

107Proposition 7. – Sensitivities of the chance P ₂:

1. The derivative instrument of the probability P ₂ with respect to the excitability ? is given by:

   

2. The derivative of the probability P ₂ with respect to the fast hoped-for return ? is given by:

   

3. The derivative of the chance P ₂ with honour to the leverage is given by:

   

4. The derived function of the probability P ₂ with respect to the pursuit rate r is apt by:

   

5. The first derivative of the probability P ₂ with respect to the sentence to maturity date t is given by:

   

Proof. – See Appendix B.

108Applying results of Proposition 7, we can prove the next properties:

109Corollary 8. – Under specific assumptions, [19] the derived function is isometric to 0 for a item value of ?₀ and the first derivative is equal to 0 for some classical value of ?. [20] In that case, the probability P₂ is an increasing so rallentando go of the instantaneous expected return ?. The probability P₂ is profit-maximizing with respect to the meter to maturity t. The derivatives and are e'er not negative.

110Test copy. – Ensure Appendix B.

111To illustrate these properties, we consider the following parameter values to assess the magnitude of the probability P ₂: ? = 8%, ? = 20%, r = 3%, t = 0.5, L = 2:

112Therefore, the probability P₂ is equal to 4.9%, and if the management period is set at one year, this probability becomes 6.82%. This feature is not negligible and makes a difference to investors. Figure 4 displays the different sensitivities of the probability P₂ .

Figure 4

Sensitivities of P₂

Sensitivities of P₂

113As mentioned in Corollary 8, the behavior of the chance P₂ as a function of the unpredictability of the stock index is rather complicated, since IT too depends on the value of the other parameters. For olive-sized values of t (typically to a lesser degree one year), it is first progressive, then reaches a localised level bes, and so becomes allargando again to reach a localised minimum in front finally decent increasing again. Nevertheless, for values of the volatility betwixt 5% and 40%, the chance is between 5.57% and 6.67%, meaning that the sensitiveness is small. The probability P₂ is accelerando in the leverage because the ratio (?a/L) is also multiplicative in the leveraging. Thus, we are computing a probability of the equal random variable on a wider interval. Moreover, IT shows a high degree of sensitivity, since it rises from 4.5% to 14.5% when the leverage goes from 2 to 5. A a function of the expected return of the underlying index number, the chance P₂ is first slightly increasing and then depreciatory. Note of hand nevertheless that the magnitude of the essence is small. For example, when the likely tax return of the risky plus is between 5% and 15%, the chance is not very thin-skinned to it and lies between 4.93% and 4.55%. This arises because expected return has an opposite effect along some parts of the starboard-script side of reflexion (23). The probability P₂ is an maximising work of the risk-free rate of interest (nigh linear). The effect of the riskless rate on the probability is stronger than that of the expected hark back. For r adequate 2%, the chance is capable 4.18% and for r capable 6%, it is adequate to 6.99%. This is because interestingness rate enters Formula (23) solely in the first cumulative distribution function. Finally, for this numerical illustration, the chance P₂ is an increasing and concave operate of time.

114We can gain more insight into the behavior of the chance P₂ as a social occasion of the excitability by considering simultaneously the result of another variable. In the following 2 figures, we suck in 3D plots where the expected restitution and the time are introduced along with the volatility. Along this cypher, we tin see that the local maximum is only obtained for very small level of the unpredictability independently of the value taken by the expected return. Furthermore, the value of this topical anaestheti maximum is strongly decreasing with the value of the expected return.

Figure 5

Sensitivities of P₂ w.r.t. two parameters

Sensitivities of P₂ w.r.t. deuce parameters

3.2 – CPPI equivalence

115In this section, we set up that, in continuous-time, a CPPI portfolio with a dump graduated to the apprais of the portfolio itself is a constant assignation portfolio strategy. First, recall that usual CPPI strategy provides portfolio indemnity and is designed to give the investor the ability to limit downside chance while allowing some participation in upside markets. Such method allows investors to recover, at maturity date, a given percent of their initial upper-case letter, in particular proposition in falling markets. The CPPI method consists of managing a high-voltage portfolio so that its value is above a floor F at any time t. The evaluate of the storey gives the dynamically insurable amount. Information technology is assumed to acquire reported to:

116

117Obviously, the initial floor F ₀ is fewer than the initial portfolio value V^CPPI ₀. This departure V^CPPI ₀ ? F ₀ is called the cushion, denoted by C ₀. Its value C_t at any time t in [0,T] is given by:

118

119Denote by e_t the exposure, which is the total amount invested in the risky asset. The standard CPPI method consists in letting e_t = mC_t where m is a constant named the multiple. The interesting case is when m dangt; 1, that is, when the payoff function is convexo-convex. Thus, present the constant dimension is 'tween the amount invested in the unsound asset and the cushion, whereas in continuant-mix strategies, this constant proportion is kept up betwixt the amount endowed in the unsound asset and the portfolio value.

120The esteem of this portfolio V^CPPI ₀ at any time t in the historical period [0,T] is given away [21]:

121

122Thus, the CPPI method is parametrized by the initial storey F ₀ and the multiple m.

123We now study the link between CPPI portfolio and constant-commingle portfolio scheme. It is usually believed that a CPPI with a zero floor and 0 danlt; m danlt; 1 is a necessary and ample condition for this scheme to be a constant mix strategy. [22] Nevertheless, we express in what follows that, at least formally, it is not necessary that the take aback is equal to zero. More precisely, in the following proposition we discuss the conditions under which these two methods are equivalent.

124Proposition 9. – A CPPI portfolio with a multiple m and with a stochastic floor F proportional to the portfolio value at any time t in the period [0,T] (proportion equal to ?) is a constant quantity-mix portfolio strategy with leverage coefficient L if and only if we have:

125

126Proof. – A floor proportional to the portfolio value at whatsoever metre t in the period of time [0,T] is written:

127

128Then, the risky exposure is written:

129

130The evolution at time t of the CPPI portfolio assess is given by

131

132which implies that:

133

134It can comprise checked that setting L = m(1 ? ?) leads Equating (26) which describes the organic evolution of the constant allocation portfolio value.

135Leveraged CPPI portfolios at inception are obtained when m dangt; 1/(1 ? ?) (i.e. e ₀ dangt; V^CPPI ₀).

136Thus, the important lineament that makes a constant proportionality strategy becoming a unceasing assignation strategy is that the floor is proportional to the portfolio value at any time. This means that, contrary to the usual CPPI, there is no thirster whatsoever downside protection because the dump is now stochastic and can thence orbit zero [23].

4 – Comparison of Leveraged ETF and Leveraged CPPI Portfolio

137In this section, we first provide a detailed analysis of the value process of the leveraged CPPI portfolio when, contrary to previous segment, the floor is not always equivalent to a fixed proportion of the portfolio value but single equal thereto at a finite telephone number of dates. This case is illustrated by the SGAM Leveraged CAC 40 scheme. It corresponds to a maximum pic to the risky asset of 200%. Then, we consider other parameter values for the multiple m and the percentage ? with a two standard values of the leverage coefficient L:L = 2 or L = 3. We examine the respective cumulative distribution functions of their returns. We besides introduce Kappa measures to compare their performances. Eventide if we assume that this strategy is rebalanced in continuous time, some features of this intersection prevent a full explicit formula for the value of the leveraged CPPI portfolio from being obtained. This is because the floor is adjusted at the beginning of every month and then held constant during the calendar month. We lead a Monte Carlo simulation in discrete time in order to equivalence LETF and leveraged CPPI portfolio.

4.1 – Leveraged CPPI

138The Leveraged CPPI fund is very similar to the CPPI funds introduced in Section 1. For instance the fund denoted LCPPI with parameter values m = 4 and ? = 50% and is designed to have level bes exposure to the risky asset of 200%. Still, this fund English hawthorn reach fewer than 200% dangerous asset photo during its lifetime. In this deference, it differs from the standard LETF, for which the leveraging is perpetual through and through time. It also has a unit of time flooring of 50% of the first-of-the-month portfolio esteem, which remains constant during the rest period of the month. The value of the multiple at the inception of the fund is set at 4.6 and subsequently allowed to vary between 3.00 and 5.50. It is readjusted at the offse of each quarter. This point is non addressed hither because it is a military science asset apportioning determination. For expositional purposes, we consider a never-ending triplex m (m = 4 for this first example) with an exposure that is arrogated to be speed bounded: it is crowned with leverage coefficient L = 2 (see suggestion below).

139Although no policy is provided for the whole full stop of management and, as a aftermath, the entire amount invested may be lost, insurance is obtained within each calendar month.

140Consider the (t_n ) _n sequence of times at which the level is resolute by the relation back [24]:

141

142Usually, t_n is the first solar day of the n ? th calendar month of the portfolio direction period. Assume that portfolio is rebalanced in discrete time during each flow [t_n ,t_{n +1}[. Let t_n,l = t_n + l.h_n . Let the rebalancing times during [t_n ,t_{n +1}[, where and h_n denotes the rebalancing frequency (typically, h_n corresponds to one day).

143Then, taking report of all previous assumptions about leveraged CPPI leads to:

144Proposition 10. – The vulnerability of the leveraged CPPI is given by:

145

146where m is the multiple and L denotes the leverage coefficient.

147The leveraged CPPI portfolio value satisfies:

148

149where the parameter L is the maximum exposure allowed for the fund [25].

150Relation (38) does not let an unequivocal result but tail be numerically illustrated by using Monte Carlo simulations (see next subsection). Nonetheless, under some additional mild assumptions, we can provide a similar-explicit solution (see Appendix C).

4.2 – Simulations and comparisons of Leveraged ETF and Leveraged CPPI Portfolios

151In this section, we implement Monte Carlo simulations to compare both strategies with a management period of unrivaled year. A discretized geometric Brownian movement is simulated connected a daily ground over a period of 240 days [26]. We use the same values of grocery store parameters as previously and 1,000,000 paths for the implicit in strain indicant are drawn. And then, a 2x LETF and a 3x LETF portfolio strategy are simulated. We consider leash Leveraged CPPI strategies delineate in Table 3:

Table 3

Leveraged CPPI Strategies

Leveraged CPPI Strategies

152All threesome strategies are designed to provide an initial exposure to the risky asset of 200% (resp. 300%) when L = 2 (resp. L = 3). Although the last case, LCPPI 2, is unrealistic, the value of the multiple being a lot too high, especially for a parentage index, information technology is utilised for expositional purposes [27]. These strategies are also simulated along the 1,000,000 paths of the stock index.

153Table 4 show the first four moments of the paying back corresponding to the iv strategies for a leverage of 2 and 3.

Tabular array 4

First four Moments of Returns of LETF and Leveraged CPPI Strategies

First four Moments of Returns of LETF and Leveraged CPPI Strategies

154For a leverage of two, the world-class two portfolio strategies arrive out as very close. Thence, the LCPPI strategy is essentially equivalent to a 2x LETF with a trifle less expected return and a trifle more downside protection, as will be highlighted in Figure 6. The LCPPI 1 and LCPPI 2 portfolios exhibit less expected returns, fewer volatility, more incontrovertible asymmetry and more extreme risk than the first two strategies.

Frame 6

CDF of the four portfolio strategies, L = 2

CDF of the quaternion portfolio strategies, L = 2

155When considering a purchase of three, the difference of opinion between the first deuce portfolios is a little snatch more remarkable.

156Figure 6 represents the additive distribution function (CDF) of the quadruplet portfolio strategies for a leverage of 2. IT highlights the downside protection that is provided by the Leveraged CPPI strategies. This issue is all the more marked as the values of the multiple m and of the story parameter ? are high.

157For instance, the probability that the return on the LETF is less than –20.64% is 24.84%, whereas the same probability for the LCPPI 2 fund is 20.72%. The LETF curve crosses the LCPPI curve at a return level of 0.794%, the LCPPI 1 curve at a return level of –1.51% and the LCPPI 2 curve at a return level of –2.900%.

158Figure 7 represents the CDF of the iv portfolio strategies for a leverage of 3. Information technology highlights the downside protective cover that is provided by the Leveraged CPPI strategies. This gist is each the more marked as the values of the multiple m and of the base parametric quantity ? are high.

Figure 7

CDF of the four portfolio strategies, L = 3

CDF of the four portfolio strategies, L = 3

159For instance, the probability that the return connected the LETF is less than -21.54% is 35.074%, whereas the Saame probability for the LCPPI 2 fund is 30.61%. The LETF curve crosses the LCPPI curve at a return level of 6.280%, the LCPPI 1 curve at a return level of 1.910% and the LCPPI 2 arc at a return level of -2.941%. It implies that the probability to bear any loss is ever higher for the LETF than for LCPPI and LCPPI 1, while the probability to bear any loss beyond -3\% is higher for the LETF than for the LCPPI 2.

160Another point needed to personify addressed is the case where the value of the bigeminal of the LCPPI strategy with ? = 50% is such that the median returns on the LETF and on the LCPPI funds are balanced. In our example, this value of the multiple denoted by m* is more or less 4.5625 (resp. 6.65) for L = 2 (resp. L = 3).. We thusly obtain the moments for the two strategies that appear in Table 5.

Table 5

First four Moments of LETF and LCPPI Strategy in the case of equal medians

First four Moments of LETF and LCPPI Strategy in the case of equal medians

161Thus for a 2x leveraging, when the medians are congeal equals, the LCPPI slimly dominates in a median-variance sense (as well as in a tight-variance sense) the LETF. Even the skewness is better for the LCPPI, just the kurtosis is worse. For the 3x leverage case, the slight median-variance (too as the stingy-variance) dominance is reversed.

162We can continue in the indistinguishable way but with equalization of expected returns. Thus, we are led to consider the case where the value of the duplex of the LCPPI strategy with ? = 50% is so much that the expected returns on the LETF and on the LCPPI funds are equal. In our example, this measure of the multiple denoted aside m* is approximately 4 (resp. 7.4005) for L = 2 (resp. L = 3). We thus obtain the moments for the two strategies that seem in Table 6.

Table 6

Number 1 four Moments of LETF and LCPPI Scheme in the case of equal means

First four Moments of LETF and LCPPI Scheme in the guinea pig of equal means

163In this eccentric, the LCPPI strategy is almost equivalent to the LETF. Actually, IT is slightly more dominant if we consider just the first trey moments.

4.3 – Z and Kappa performance measures of Leveraged ETF and Leveraged CPPI Portfolios

164In this section, we compare leveraged ETF and leveraged CPPI strategies by means of the Omega performance measure. Indeed, some strategies are non linear in the risky asset due to the leverage characteristic. This induces asymmetric returns and higher probability of red. Thence, downside risk measures are required for such products. The Omega and the Kappa functioning measures are supported such risk measures. [28]

165The Omega public presentation measure has been introduced by Keating and Shadwick (2002) and Cascon et al. (2003) to take Sir Thomas More account of the likely losses than standard performance measures such Eastern Samoa the Sharpe ratio based only on the mean and the variance of the dispersion of the returns. Z measure involves the entire return distribution by separating the returns below and in a higher place a fixed loss brink since Omega is equal to the chance weighted ratio of gains to losses relative to a return verge q:

166

167where F(.) is the cumulative distribution function of the plus return defined happening the interval (a,b), with respect to the probability distribution ? and q is the reappearance doorstep q selected by the investor. Returns below the loss doorstep are viewed as losses and returns supra as gains. For a fixed threshold, investor should always favour the portfolio with the highest Omega value.

168As established aside Kazemi et al. (2004), Omega can be handwritten as:

169

170Thus Omega can make up well-advised as the ratio of the prices of a phone option to a put option written on X with strike price q just some evaluated under the historical probability ?.

171Kazemi et al. (2004) preface the Sharpe Omega criterion defined by:

172

173The Kappa measures introduced by Kaplan and Knowles (2004) are defined by:

174

175For l = 1, we acquire the Sharpe Omega measure and, for l = 2, we regai the Sortino ratio. Note also that Zakamouline (2010) shows that Kappa measures are performance measures based on piecewise linear summation mightiness utility functions. [29]

176By means of Four-card monte Carlo experiments, we simulate the Omega of a 2x and a 3x leveraged ETF A well as the Omega of the corresponding leveraged CPPI strategies; as in the previous incision. We apply the same set of parameters for the commercial enterprise market.

177In Table 7, we report the results of the simulations [30]. We study the effect of the threshold connected the ranking of the strategies reported to the Z touchstone. We find again confirmation that the leveraged CPPI strategy offers to investors more downside protection that LETF. Indeed, it is for low level of the threshold that LCPPI strategies have higher Omega values. Again, this gist is all the much well-marked as the values of the multiple m and of the storey parametric quantity ? are piercing. Thus, it is when investors are many concerned about protecting their cap that LCPPI strategies show higher Z values than LETF strategy. Notice that the LCPPI 2 is never a dominant scheme. In one case the threshold value is higher, LETF is predominant with respect to Omega. Note though that LETF and LCPPI with ? = 50% are very close.

Table 7

OMEGA of LETF and LCPPI Strategies for L = 2 and L = 3

OMEGA of LETF and LCPPI Strategies for L = 2 and L = 3

178We consider now the Kappa criterion for l = 2, a Sortino type ratio. Table 8 reports the results of the simulations. Compared with Omega results, we find that Kappa 2 quantity [31] leaves near unchanged the superior of the different strategies.

Table 8

KAPPA 2 of LETF and LCPPI Strategies for L = 2 and L = 3

KAPPA 2 of LETF and LCPPI Strategies for L = 2 and L = 3

5 – Closing remarks

179In this theme, we analyse main properties of leveraged ETFs that correspond to convex constant apportionment portfolio strategies. In particular, we illustrate how the stock market index throne experience an increase piece at the same time the leveraged fund decreases, calculating the explicit chance of so much an event for the GBM lawsuit. We also establish that, in continuous-clip, a CPPI portfolio with a floor progressive to the value of the portfolio itself is also a convex constant apportioning portfolio scheme. Taking a leveraged ETF issued past SGAM and managed under a somewhat altered CPPI scheme that we call Leveraged CPPI, we aim a quasi stated expression for its value. We compare both strategies by way of their first four moments as well as past applying Omega and Kappa performance measures. These measures are especially healthy designed for portfolios with asymmetric returns. Monte Carlo simulations allow us to level out the similarities in behavior of a standard leveraged ETF and of a Leveraged CPPI strategy. Nevertheless, we can conclude that the Leveraged CPPI strategy is essentially equivalent to a LETF with a little less expected return and a weeny many downside aegis.

Appendix A – Static leveraged strategy

180Under supposition (12) on the risky plus dynamics and with discrete-meter rebalancing, we derive:

Thus, for the GBM case, the ergodic variable has a Lognormal distribution . The probability denseness function of is planned in Figure 8 for three different levels of leverage (L = 2, 3 and 5) and the same parametric quantity values as in the main text are used:

Cecal appendage B – Sensitivities of the chance P ₂ that the stock index increases spell the leveraged fund decreases

181Proof of Proposition 7

182To determine the sensitivities of P ₂, note that we have:

183

184i) Derivative of the probability P ₂ with respect to the volatility ?. We have:

185

186ii) Derivative of the probability P ₂ with respect to instantaneous expected pass ?:

187Since , we bear:

188 , from which we deduce the result.

189iii) Determination of the derivative of the probability P ₂ with respect to the leverage L:

190We have:

191

192Then, from relation we deduce the result.

193iv) Derivative of the probability P₂ with respect to the rate of interest r:

194We have:

195

196Then, using , we get the result.

197v) Derivative of the probability P₂ with respect to the time to maturity t:

198We have:

199

200Therefore, victimisation , the result is proved.

201Proof of Corollary 8

202i) Condition is satisfied if and only if we have:

203

204Since L dangt; 1, the last term of this par is non negative. Thus, a solution ? ₀ can exist as shown in numerical example.

205cardinal) Note that, since L dangt; 1 we have a danlt; 0. Additionally, since , we get the following equivalence:

206

207This leads to the condition: Thus,

208

209Finally, we acquire a specify which corresponds to:

210

211Note that this condition does not depend on fourth dimension to maturity t. As shown in next denotative example, the derivative bottom be equal to 0 for standard value of ? (in the example, ? ?4.9%). It means that the probability P ₂ can be an increasing then tapering function of the instantaneous expected return ?.

212trey) dangt; 0

213iv) dangt; 0 since L dangt; 1

214v) If (regulation case), a sufficient circumstance to get is that

215

216This leads to a second-order mathematical function equation with respect to L:

217

218with a discriminant . Thus, an interior solution give the sack exist but the only supportive solution is (for the standard case ? dangt; r). The condition L ₁ dangt; 1 is tantamount to . For the future numerical example, L ₁ = 3. Thus, since we consider the case L = 2, the probability P ₂ is multiplicative with respect to the clock to maturity t.

Appendix C – Quasi-explicit root for Leveraged CPPI fund

219Assume now that the portfolio director trades in continuous-time during for each one period [t_n ,t_{n +1}[. Suppose besides that the CPPI scheme is not capped. So, the portfolio value can be determined by induction as follows:

220Recall that the standardized CPPI scheme, with a deterministic floor F_t , induces a portfolio esteem defined by:

221

222where ?_t = C ₀ exp [?_t ] and .

223Consequently, applying relation (42) for each management period [t_n ,t_{n +1}[, we fundament determine the portfolio time value of the leveraged CPPI:

224Denote for all s danlt; t,

225

226We have:

227

228

229Then, we deduce: If the leveraged CPPI is continuously rebalanced and non capped, then its value is donated by: for any t ? [t_n ,t_{n +1}[,

230

231Consider for case the geometric Brownian example. Simulate also to simplify that durations (t_{n +1} ? t_n ) are equal to a constant ? and that portfolio maturity T is equal to t_N . Then, the ratio is given past:

232

233i) The portfolio measure V_T has a similar-explicit form:

234

235Note that the value of return on the historical period [t_n ,t_{n +1}[is given by:

236

237ii) The probability statistical distribution of portfolio return can be determined as follows: The random variables are i.i.d. with Lognormal distribution. So, the logreturn of the portfolio ln (V_T/V ₀) is the union of N i.i.d. random variables. Their coarse probability distribution is an affine transformation of the Lognormal distribution with parameters . Thus, the Pdf of the logreturn of the portfolio ln (V_T/V ₀) is given by [32]:

238

239where is given away:

240

Notes

• [*]

  CERGAM, Aix-Marseille University, Aix-Marseille School of Economics, Aix-nut-Provence, France and Euromed Management, Marseille, France.
  Teldannbsp;: 33 (0)4 91 14 07 43. E-mail: philippe.bertrand@univ-amu.francium

• [**]

  THEMA, University of Cergy-Pontoise, 33, Bd du Port, 95011, Cergy-Pontoise, France Teldannbsp;: 331 34 25 61 72; Telefax: 331 34 25 62 33. Email: jean-luc.prigent@u-cergy.fr
  The authors thank the anonymous referee for the scrupulous reading and the valued comments and suggestions to ameliorate this paper. We would also like-minded to thank Patrick Roger for his face-saving comments as well American Samoa participants at the 28th annual International Conference of the French Finance Association held in Montpellier, 2011. We are also grateful to participants at the 28th time period coition "Journées de Microéconomue appliquée", (Susah, JMA 2011). We also thank participants to the PRMIA seminar in Munich, January 2012. The regular disclaimer applies.

• [1]

  In 2009, the full amount of leveraged ETFs traded in the The States commercial enterprise market was higher than US$31 billions.

• [2]

  SGAM stands for "Société Générale Asset Management". AI means "Alternative Investment".

• [3]

  IT likewise publishes short, double unretentive, triple squab and triple leverage indices.

• [4]

  To the champion of our knowledge, leveraged funds have not been analyzed as CPPI portfolio in the U.S.A.

• [5]

  After a reorganization of the asset management activity of Société Générale, these funds are now managed by LYXXOR since the 1st of September 2009.

• [6]

  See the "prospectus complet" of the SGAM ETF Leveraged CAC 40, certified by the "Autorité diethylstilboestrol marchés financiers" (AMF).

• [7]

  This latter indefinite is illustrated by using the SGAM Leveraged CAC 40 scheme.

• [8]

  See Appendix A where the nonmoving leverage case is discussed for the geometric Brownian motion case.

• [9]

  See Perold (1986), Black and Perold (1992) and Bertrand and Prigent (2003, 2005) for more details about the CPPI scheme.

• [10]

  For expositional simpleness and w.l.o.g., we assume here that rate of interest is capable zero in.

• [11]

  As agelong arsenic we debate an absolute chronological succession of returns.

• [12]

  See below Section (3.1) for a formal statement of this property.

• [13]

  The sensitiveness to the volatility (the Vega) is negative for CPPI portfolios atomic number 3 shown in Bookstaber and Langsam (2001) and Bertrand and Prigent (2005).

• [14]

  In the context of use of the naturally occurring article, the stock market index is the ETF. Additionally, we seize that all Price processes are dividend-adjusted.

• [15]

  See also Cheng and Madhavan (2009) for the GBM encase.

• [16]

  The showcase, L danlt; 0, a bear or reverse leveraged scheme, can be treated similarly.

• [17]

  Instead, the firstborn two components fire be organized in the following direction:

  • Leveraged return along the index,
  • Financing toll, r (1 ? L) danlt; 0 for L dangt; 1.

  The volatility correction is the same.

• [18]

  In discrete-time, this probability corresponds to

  

• [19]

  See Appendix B and next nonverbal example.

• [20]

  In the example, ? ? 3.8dannbsp;%.

• [21]

  This formula was proven aside Opprobrious and Perold (1987) and has been extended to the stochastic volatility case by Bertand and Prigent (2003).

• [22]

  See Escobar et al. (2009).

• [23]

  We could have set the treasure of the floor such that: F_t = ?V^CPPI ₀ ? F ₀, 0 danlt; ? danlt; 1. In this case, we would obtain a constant-shuffle-type scheme in which the part of the wealthiness invested in the risky asset is no longer proportional to the unscathed wealth at each time, but rather to (V ^CPPI _t ? F ₀).

• [24]

  The parameter ? = 50% for instance.

• [25]

  In the case of the SGAM product, IT was set at 200%.

• [26]

  Here a conventionalized year is assumed to contain twelve months of 20 days.

• [27]

  Notice that for L = 3, the case m = 12, the case is too unrealistic.

• [28]

  We concern to Bertrand and Prigent (2011) for an in-depth comparability of casebook portfolio insurance strategies with the Omega and the Kappa measures.

• [29]

  To prove this upshot, introduce the following utility social function:

  

  where ? dangt; 1 and n a nonzero integer. As proved away Zakamouline (2010), the standard investor's capital parceling job yields to the following relation:

  

  where E[U*(V)] denotes the utility of the optimal allocation. Thus, E[U*(V)] is an increasing shift of the Kappa (n) ratio, which proves that this latter one is based on the utility U_q . Note also that U_l is convex connected]??,q] if and only if we have n = 1. This corresponds to the Omega measure, obtained at the limit when n ?1.. Thusly, the Omega functioning meter is accompanying the maximation of an likely utility with passing aversion, as introduced by Tversky and Kahneman (1992).

• [30]

  The highest value of Omega for from each one threshold value is in bold.

• [31]

  We perform the same simulations with a Kappa 3 measure. The ranking is the same arsenic that obtained with Kappa 2 measure which is the Sortino ratio.

• [32]

  The symbolic representation * denotes the convolution intersection.

position based trading strategy for leveraged etf

Source: https://www.cairn.info/revue-finance-2013-1-page-73.htm

Posted by: bitneralonds.blogspot.com

Share this post