20160505(목)_다음논문

1. Tobias Moskowitz_Chicago Booth
 http://faculty.chicagobooth.edu/tobias.moskowitz/research/chronology.html#2014

2. Andrea Frazzini_ AQR
http://www.econ.yale.edu/~af227/

Value and Momentum Everywhere by Clifford S. Asness

1. We offer new insights into these two market anomalies by examining their returns jointly across eight diverse markets and asset classes.

2. Provide several key questions about these pervasive market phenomena

• How much variation exists in value and momentum premia across markets and asset classes?
• How correlated are value and momentum returns across these diverse markets and asset classes with different geographines, structures, investor types, and securities?
• What are the economic drivers of value and momentum premia and their correlation structure?
• What is a natural benchmark model for portfolios of global securities across different asset classes?

<What they found>

1. Consistent and ubiquitous evidence of value and momentum return premia across all the markets we study, including

• Government bonds : Value and momentum
• Currencies and commodities : Value

2. They found significant comovement in value and momentum strategies across diverse asset classes.

• Value strategies are positively correlated with other value strategies across otherwise unrelated markets
• Momentum strategies across diverse asset classes
• Value and Momentum are negatively correlated with each other within and across asset classes.

3. The striking comovement pattern across asset classes is one of our central findings and suggests the presence of common global factors related to value and momentum.

...

4. We show that separate factors for value and momentum best explain the data, rather than a single factor, since both strategies produce positive returns on average yet are negatively correlated.

5. We investigate the source of this common global factor structures.

• Modest link - Macroeconomic factor
• Significant evidence - liquidity risk(negatively related with value and positively related with momentum) ==> Value 와 Momentum의 negative correlation 의 일부분은 여기서 나온다고 생각할 수 있다.
• Funding risk - Importance of it increase more and more. 
• 이 결과들은 gained by looking across many markets at once

6. 사실 liquidity risk 는 많은 부분을 설명하지는 못한다. 특히 Value 와 Momentu을 합쳤을때의 우수한 결과는 liquidity risk 가 희석된 상태. ==> Hence, funding liquidity risk can only provide a partial and incomplete explanation for momentum, but cannot explain the value premium or the value and momentum combination premium.

7. 우리가 밝은 단서들이 value and momentum premia의 존재를 설명하는 단서가 된다. 

• 예를들어 두 팩터의 강한 음의 상관관계는 두 팩터를 모두 설명할 수 있는 하나의 리스크 요인이 존재할 수 있음을 이야기 해주는 단서가 된다. 
• 반대로 이러한 상관관계는 behavioral model의 존재에 대한 단서는 아니다.

8. 여러 자산들을 통해 발견한 value 와 momentum 에 대한 결과로 인해 rational asset pricing 이론들로(firm investment risk or firm growth options을 이용한) Value 와 Momentum 에 관한 프리미엄을 설명하는데는 무리가 있음을 알 수 있다. currencies, government bonds, commodities에서 발견한 것들을 설명할 수 없다.

9. 논문에서 Value와 Momentum 을 측정하는 방법 

• Stock : BE / ME (book to market ratio)
• Book Value의 경우 6개월 lagged to ensure data availability
• Momentum : 과거 12개월 return
• 최근 1개월은 skip which is standard in the momentum literature to avoid the 1-month reversal in stock returns -> 관련논문 참고, liquidity issue 가 있다는데...??
• 다른 asset classes 
• Momentum : 다른 asset에서는 최근 1개월을 삭제 하지 않아도 됨(liquidity issue가 적음) 또한 최근 1개월 포함할때 성과 좋음. 하지만 이 논문에서는 equity랑 똑같이 맞춰줌(1개월 삭제) , 보수적
• Value  : 주식같은 value 가 없음. 따라서 모든 자산 class에 같은 기준을 적용하기 힘들다. 
• country indices : 전 월의 BE/ME ratio for the MSCI index of the country
• 기타 다른 자산 : log of the spot price 5 years ago ( actually, the average spot price from 4.5 to 5.5 years ago) divided by the most recent spot price
• 이렇게 하는게 통계적으로 의미가 있음. 관련 논문 참고.

10. 포트폴리오 구성은 3분위로 구성

• 주식 : 시총으로 비율
• 기타 : equal weight

11. Value and Momentum Factors
각각은 HIgh - low 로 구성. 비율의 경우 개별 rank - cross-sectional average rank of that signal.

12. 벤치마크 설정

• 개별주식 : 각각 나라 MSCI index
• Country index : world index
• 나머지 자산 : equal-weighted basket of the securities in each asset class

<결론>
1. 모든 주식시장에서 Value 주식의 선전. Japan에서 특히 가장 강
2. 모멘텀 또한 강함. 특히 유럽에서. 일본에서 통계적 유의성이 좀 약함.
3. Value 와  Momentum의 상관관계가 대략 -0.6 정도. 둘다 positive series이기 때문에 둘을 섞으면 매우 좋아질 거라고 예상 가능
4. 일본시장에서 사실 밸류가 상대적으로 매우 강해서 모멘텀이 약한데(코릴은 낮기 때문에) 이런 일본 시장에서도 벨류와 모멘텀을 섞으면 샤프가 더 좋아짐
5. 개별 국가들을 섞어 global portfolio를 만들었을때 샤프가 크게 좋아지지 않는 모습을 볼때, Value 는 Value끼리 Corr 이 높고 Mom은 Mom끼리 Corr이 높다고 생각 가능
6. 모든 자산에 대해서도 포트폴리오를 구성해 봤다.

• 개별 자산의 vol 차이가 매우 크기 때문에 1/vol 로 weight을 줘서 계산 해 봤는데 샤프가 그다지 향상되지 않았다.
• 즉 모든 자산간에 해당 팩터들에서 상관성이 존재 한다는 이야기

20160424 - 현재 매크로 상황 정리

시장을 크게 네가지 국면으로 나눠 보자. 2x2.

리스크 온
리스크 오프
x
Us Treasury yield High
Us Treasury yield Low

2016.1월 초 : 12월 FOMC, 금리 올림. 과거에 질러놓은 이야기 때문에 시장의 기대(상승한 금리)에 부합하는 행동이 차라리 행동을 하지 않는 것보다 시장에 주는 충격에 덜 할 것이라고 생각. 금리 인상. 중국+이머징 작살. 현재와 같은 상항에서 금리를 올리면 어떻게 되는지 알게 됨. 앞으로 쉽게 금리를 올리지는 못 할 것.
===> Risk Off &  Us Treasury Yield headed to Low

2016. 02.11 ~ 2016.03 중순 : 시장이 지나치게 빠졌던 상황에서 반등. (2월 11일 S&P 저점. Double bottom). 건강한 강세장의 모습 이었음.
===> Risk On & Us Treasury Yield headed to High

2016.03 중순 ~ 20160410정도? : 201603 중순을 기점으로 주가 상승 & 금리 상승이 함께 있는 상황은 종료 됨.  S&P 입장에서 현재 레벨을 돌파하기 위해서는 또 다른 에너지가 필요한 부담 스러운 레벨 이었는데 중앙은행 발 달러 약세로 인해 한 레벨 더 오를 수 있었음. 금리 안정적, 유가 상승, 주가 상승.
===> Risk On & Us Treasury Yield Headed to Low

20160410 ~ 현재 : 지속적으로 Risk On. Risk On의 관성이 남아 있는 시점. Risk On 을 원해서 모든걸 Risk On 의 신호로 해석하고 있는 단계. 다만 지금 눈여겨 볼 점은 금리가 상승하기 시작 했다는 것.
===>  Risk On & Us Treasury Yield has started to go High


현재 : 달러 약세의 바닥을 테스트 하는 지점. 달러 약세가 강세로 추세 전환 하기 위해서는 에너지가 필요. 어떤 에너지가 있을까?
모건스탠리 - 20160415 보고서

Additionally, our FX team notes the following triggers for a resumption of USD strength: i) Weaker EM data and in particular in China, where our base case is that this happens in 2H16 as the fiscal stimulus fades; 
ii) Additional easing measures by DM central banks to fight the USD weakness, notably Japan; --> 4월 28일 BOJ 
iii) Continued closing of the US output gap, sparking questions about how long the Fed can remain on hold;
 iv) Commodity prices reversing the recent rally. Specifically on oil, our oil strategists believe disappointment will come from non-compliance with any potential deal. 

Finally, we also think core rates represent a important risk due to the very low levels, particularly given that markets are now reaching lows similar to a year ago, which were indeed followed by a sharp sell-off. 

2. 일본
최근 옌이 강했는데 니케이가 올랐다.!


3. 중국

4. 유럽

5. 한국
- 원달러 : 주요 이평선을 다 깨고 내려 왔는데 굳이 떨어지는 칼날을 받을 필요가 있는가 싶다.
- 채권 : 1.43% 찍고 현재 1.47%

수익률 계산

1. Equity 포트폴리오를 총 얼마 매입 했는지를 계산 (첫날 매입 금액)

2. 선물 포트폴리오를 매입 하는데 비용이 얼만큼 들었는지 계산

3. 1+2 를 총 사용 금액으로 결정
  - 현재는 현물 7 : 선물 1 의 비율로 사용 금액을 결정. 즉 3번 고정.


4. Daily 수익률 계산

1. Daily equity 누적 수익률 계산 = Equity 누적 벌어드린 금액 총 합계 / 3
1. Daily equity 당일 수익률 계산 = 1 / t-1일 누적
2. 참고 : 분모를 선물 사는데 필요한 돈으로 수정 해주기 전 pure 한 수익률도 표시해 준다.
2. Daily 선물 누적 수익률 계산 = 선물 누적 벌어드린 금액 총 합계 / 3
1. 선물 누적 벌어드린 총 합계 계산 시, 

	코스닥매입금	계약수	종가
20160401	600	-10	600
20160402	605	-3	603
20160403	607	-1	605
20160404	604	1	604
20160405			605

이런 정보를 바탕으로 당일 까지 누적 금액 계산하기
==> daily 수익률은 t일까지 누적 수익률 /t-1 일까지의 누적 수익률


5. Daily




**** Equity 포트가 오름으로써 생길 수 있는 비용에 관한 문제
1. 선물 가격이 떨어 지면서 추가 납입금이 늘어 날 수 있음
2. Equity 포트의 notional 이 커짐으로써 Short 을 해줘야 하는 계약수가 늘어남

CH15 : Performance Measurement and Attribution

Performance attribution dissects the return in order to pinpoint the exact sources of value

Important function of performance measurement is to determine whether the portfolio manager outperformed or underperformed the benchmark and whether the difference was due to skill or luck.

Source of excess returns is key to understanding how well and efficiently the factor model and all other parts of the investing strategy are working.


15.2 Measuring returns

CH3 : Basic QEPM Models

3.2 Basic QEPM models and portfolio construction procedures

There are two generic factor models that are used to determine how stock returns and risks vary with factors

• Fundamental factor model
• Economic factor model

They employ different techniques for modeling stock returns.

Average stock return is determined by the product of the factor premium and the factor exposure.

• Factor premium : How much investors are willing to pay for each factor 
• Fundamental factor model : Must be estimated with historical relationships between stock returns and factor exposure
• Economic factor medel : Can be determined up to a propotionality without a statical estimation in certain cases
• Factor exposure : measures how sensitive the stock return is to a factor
• Fundamental factor model (ex> p/e, marketcap..) : directly observerable 
• Economic factor model  : Must be estimeated

<Steps>

3.2.1 Factor Choice

3.2.2 The data decision

• Cross-sectional dimension : The cross-sectional dimension defines a data set by the characteristics of the stocks it includes 
• 특정 data set 의 특징은  최종 포트폴리오에 영향을 미칠 것이다.
• 특정 data set 에 속한 주식의 숫자는 estimation을 쉽게 만들 것이다.
• Time-series dimension : 

3.2.3 Factor Exposure

• Fundamental factor model : easy
• Economic factor model : 뒤에서 이야기 하겠지만 먼저 factor premium을 계산하고, 이 factor premium과 return과의 관계를 통해서 계산한다.

3.2.4 Factor Premium

20160209_Some articles

1. 5 Things You Have To Learn From Kyle Bass

http://blog.stocktwits.com/5-things-we-learned-from-kyle-bass/

2. 'Risk Parity' strategy shows strain
http://www.ft.com/intl/cms/s/0/8542e88e-cea6-11e5-92a1-c5e23ef99c77.html#axzz3zfwok1Cd

Last year was a terrible one for “risk parity”, once one of the hottest strategies in the investment world, as losses mounted and some analysts blamed it for exacerbating market turbulence. So far 2016 has offered little respite.

...

But the performance started to sag in 2014, and nosedived last year, ruining its reputation for resilience in almost any conditions. The Salient Risk Parity index slumped 12 per cent in 2015 and JPMorgan’s gauge of risk parity fund performance fell more than 8 per cent.

...

Even pioneers of the strategy have suffered a bad time. Bridgewater’s “All Weather” risk parity fund lost 7.7 per cent last year and is down about 10 per cent net of fees since July 2014, according to a person familiar with the matter.

....

Adding to the woes(Commodity), bonds — a big part of risk parity portfolios — have failed to act as a suitable counterweight for the once seemingly bulletproof strategy.

....

“Rates have historically done well in environments like this, but last year we got nothing out of it,” says Mr Knight....“This is a world where there are some purists, and the theory underpinning risk parity is strong, so there aren’t that many that change their approach,” he says. “I don’t consider myself a purist, and there are some regimes that aren’t that risk parity friendly.”

...

The crucial question is whether this bout of underperformance is an unpleasantly long but still temporary blip, the start of a new risk parity-unfriendly market regime, or even a symptom of something more broadly awry in the fundamental assumptions underpinning the strategy.

Industry insiders are still confident it is the first, and are hopeful that a turnround beckons. Bridgewater’s All Weather fund is still “a much better ‘sleep at a night’ portfolio than any single asset class and most portfolios”, argues an executive at the hedge fund.

Similarly, Mr Mendelson at AQR says that apart from some minor implementation tweaks he has not changed anything as a result of market movements. “Systematic investing does best when you run a consistent process that doesn’t change in response to short-term fluctuations,” he argues.

3. 스마트베타, 패시브와 액티브 모두의 대안
http://m.thebell.co.kr/m/newsview.asp?svccode=00&newskey=201602010100002350000145

4. Hedgeable
https://www.hedgeable.com/hedgeable-investment-philosophy-white-paper#capm

...

Eliminating 75% of the losses during these two periods has an immense effect on portfolio growth, adding more than $1 million over 25 years. For a hard-working American investing for retirement, that extra $1 million can significantly boost their standard of living.

MPT based index investing will never achieve this kind of loss reduction; an MPT philosophy inherently accepts all losses and thus represents a huge opportunity cost for long-term investors.

...

Contrary to what the MPT zealots have made the public believe, managers have been able to achieve the kind of loss reduction shown above. Consider the tangible example of the performance of Bridgewater Associates’ Pure Alpha Fund (the firm where Hedgeable CEO Mike Kane began his career) versus the S&P:

Bridgewater produced a return that nearly doubled that of the S&P with about half the volatility, and a (-14.18%) max drawdown, thus cutting out 75.03% of the losses. Bridgewater is an alternative manager that invests in futures, options, and other vehicles that are typically only purchased for wealthy or sophisticated investors. But what if we could echo this sort of risk-managed profile by using large, liquid ETFs and stocks instead of complicated ultra-high-net-worth instruments? Later in this paper, we will explore how Hedgeable has disrupted the market by making that possible and available to everyone.

...

5. ‘Robo-advisers’ try to calm investor nerves
http://www.ft.com/intl/cms/s/0/3a44af14-c6ec-11e5-808f-8231cd71622e.html#axzz3zfwok1Cd

Daniel Egan, director of behavioural finance and investing at New York-based Betterment, said the firm had received higher volumes of calls and emails in recent weeks, and was trying to ease concerns by providing “very personalised, very positive” notifications when customers log in.

6. Which robos got off to the fastest start in 2016?

http://www.investmentnews.com/article/20160115/FREE/160119956/which-robos-got-off-to-the-fastest-start-in-2016

 

2016 1월 현재 로보어드바이져 AUM 2015 01과 비교