Skip to content

Some Thoughts On The Demise of Goldman Sachs’ “Global Alpha”

September 17, 2011

I read an article that Goldman Sachs’ flagship “Global Alpha”  fund http://www.cnbc.com/id/44545789  was closed today following some hefty losses this year.  To many it is a dark day for quantitative trading/investing when the so-called “best and brightest” are hanging up their gloves. I wouldn’t read too much into this because Goldman likely saves its best and brightest for trading its own money instead of for clients. On that note, there is a whiff of fumes from a statement by GS that revealed ” potential liabilities” to now bankrupt Lehman Brothers. With respect to their claims against Lehman they wrote: “we cannot predict when the fund will be able to make final distributions, if any, to investors in the fund.”To me this raises questions as to whether the fund served merely as a vehicle to engage in shady and unfair dealings with its own trading desks. Perhaps that is why Blankenfein recently took the unusual step for a CEO to hire a personal lawyer?

Furthermore to say that the Global Alpha fund was a high-speed computerized fund is a complete misnomer. Such behemoth hedge funds that trade for outside investors almost always run strategies that appear somewhat “conventional” (tactical asset allocation, buying undervalued securities, engaging in complicated over the counter derivative and fixed income strategies) and have modest turnover that is enhanced high speed execution.  In contrast, the strategies run internally are almost certainly higher frequency and built on tremendous structural edges gained from trading for and on behalf of clients.   To me, the closure of “Global Alpha”  highlights several points: 1) the markets are in the midst of tremendous global financial mess where conventional historical relationships have been severely disrupted by highly unpredictable movements in different assets—this was probably one of the reasons why Goldman had trouble especially since it is trading the Global Alpha with a lower frequency of trading 2) it is no longer  desirable for economic and practical reasons to trade scalable quantitative strategies on behalf of investors— their capital is not truly required for a firm like Goldman and investors are disruptive by exposing them to greater legal risk and unwanted transparency 3) the best talent (aside from the internal GS trading group) is rapidly shifting to proprietary trading and high-frequency firms where it is most beneficial.

Things are changing in the investment industry, and something tells me that this is a historic event that many will point to in the future as a turning point. The  jury is still out yet as to what these changes will be, but my intuition suggests that we are entering a new and different era for investment management. True “alpha” is quickly becoming a privatized business.

Advertisements
20 Comments leave one →
  1. Yaba Qi permalink
    September 17, 2011 3:46 am

    Well said!

    Quote: “To me this raises questions as to whether the fund served merely as a vehicle to engage in shady and unfair dealings with its own trading desks.”
    Same thoughts from my side. In 2008, Goldman was net buyer of subprime papers in the market but it showed a negative position in its books at the end of the year. The only explanation according to me: Goldman stuffed clients’ accounts and funds with the toxic waste.

    Also, when you read about Global Alpha, you realise that the fund was invested in several “risk premia”. Well, its name should have been Global Beta then.
    🙂

    • david varadi permalink*
      September 19, 2011 11:23 pm

      thanks yaba– I completely agree on the toxic waste part. I also believe that to the extent it held passive exposure to such risk premia it was taking on beta, otherwise it is possible that it was tactically allocating among such risk premia in which case it was a combination of alpha and beta (in this case negative alpha).
      best
      dv

  2. September 17, 2011 8:11 am

    Global Alpha failed because of the lack of incentives. Cliff Asness, founder of Global Alpha, left precisely because it’s not worth it when Goldman takes a huge chunk of the performance fees. Global Alpha is a case study of a failed business where lack of incentives led to thinning talent and dwindling performance.

    • david varadi permalink*
      September 19, 2011 11:25 pm

      hi henry, you are correct to an extent that this was one of the causes which I mentioned to some extent in terms of the migration of talent to other areas of the industry. it is impossible to pin this as the exact and only cause— there was a tremendous failure of incentives also internally to save the best ideas and trades for the alpha fund versus the trading desk.
      best
      dv

  3. John McClure permalink
    September 17, 2011 9:06 am

    Good article David. I hope you are well.

    John

    • david varadi permalink*
      September 19, 2011 11:26 pm

      hi john, thank you, and good to hear from you. hope you are well also.
      best
      david

  4. Damian permalink
    September 17, 2011 1:14 pm

    I think you’re dead-on here David – the only thing I would add is that Global Alpha is an example of what I call “beta + leverage” – meaning I’m not sure they were generating alpha so much as generating beta and then producing alpha via leverage.

    • Thierry permalink
      September 17, 2011 5:17 pm

      “ALPHA” is ALWAYS beta + leverage…. just think about it for 2 minutes. You always have exposure to something, whatever for a microsecond or for a year, whatever long only or long-short…

      • Damian permalink
        September 17, 2011 9:00 pm

        I don’t agree – alpha is the return above a benchmark. If you have an active, long-only manager selecting stocks from the S&P500 and he or she manages on beating the index, they have generated a positive alpha without using any leverage. Maybe you mean leverage in a different way?

      • September 19, 2011 12:08 am

        I think you may be mistaken.
        If you think about it, imagine a time series of price: Dow for instance, from t0 to t2 price rises 5%, and from t0 to t1 price falls 2%. Alpha means generating returns above the 5% at t2 (the 5% is beta), which entails capturing the theoretical 7% if long only, or max of 9% if always in.

      • david varadi permalink*
        September 19, 2011 11:30 pm

        hi thierry, you bring up a point worth expanding on in a separate post and there are definitely subtleties involved. it is not as concrete as most people imagine.
        best
        dv

    • david varadi permalink*
      September 19, 2011 11:28 pm

      Hi Damian, thank you and good to hear from you. I agree that alpha transport was definitely a factor here after reading some articles on their overall strategy. In this case the alpha was negative and the beta also lost which served to substantially increase their losses.
      best
      david

  5. Chris permalink
    September 18, 2011 5:05 pm

    I fail to see the significance of GS closing the fund. It hasn’t performed well for at least 4 years. It’s far from the flagship product that the media has made it out to be. GS has closed about 30 different funds this year; the fact that this one was once a star, prior to its top quants leaving, doesn’t mean all that much. Ask GSIP investors what Goldmans’ new flagship product is… they’ll be sure to let you know (400% return inside a few years).

    • david varadi permalink*
      September 19, 2011 11:32 pm

      hi chris, I think that the significance is more of a factor in terms of the future of asset management under an investment banking umbrella or any other situation in which there are substantial conflicts of interest. I believe new regulations may follow since GS is so large and significant according to the media and the public. perception is often more important than truth. I agree with your last point.
      best
      david

  6. Thierry permalink
    September 19, 2011 11:57 am

    Damian :
    I don’t agree – alpha is the return above a benchmark. If you have an active, long-only manager selecting stocks from the S&P500 and he or she manages on beating the index, they have generated a positive alpha without using any leverage. Maybe you mean leverage in a different way?

    So for you alpha is part of the return unexplained by the benchmark beta? What about other “benchmarks” which we usually call “risk premia”?

    Let’s say I beat a 5-10y US notes benchmark by leveraging up on an equity investment. Is this “alpha”? No it is absolutely not. It is leveraged beta! You ALWAYS have an exposure to beta premiums, even if it is just for a micro-second.

    I can be “alpha”, which means “have part of my returns unexplained by a certain set of risk premiums”, but it can very well be that I have made 250 trillions trade over one day or one month (whatever) and the aggregated result is just something you cannot regress conveniently. Hence I am alpha, but I know, I am actually beta, all the time!

    You may understand it better with market timing. If today’s markets have been going up 5% for the first part of the day and then down 3% for the remainder of the day (may be something we should get used to for the future…), market return (or call it benchmark return if you wish) is 1.85%. If I have been very good at market timing, went long the first part of the day, and short the second part of the day, then my return is 8.15%. Is this an “alpha” of 6.3% with a beta=1? No it is simply a LEVERAGED BETA investment! Beta because I was purely exposed to the market and leveraged because I changed my position during the day, probably doubling my risk: had I been wrong and shorted the first part of the day while being long the second part of the day, I would end up losing -7.85%…. Is my “alpha” in that case -9.7% with a beta=1?

    Hope this helps.

    • Chris permalink
      September 19, 2011 10:38 pm

      What on earth are you talking about? Alpha = returns in excess of a benchmark with no added risk. Period.

      • Thierry permalink
        September 20, 2011 3:57 am

        Clearly wrong definition of alpha. To keep it simple, alpha = unexplained returns, which is:

        r_t = \alpha_t + \sum_{i=1}^N \beta_i r^i_t + \epsilon_t where \epsilon_t is some kind of white noise component.

        The usual classifications of return streams in fund of hedge funds is the following:

        1) beta streams (easy exposures : equity, FI, …)
        2) alternative beta streams (less common return streams: commos, FX, insurance, ABL, etc…)
        3) alpha… or “I do not know where your performance comes from, hence I will not invest”.

        The alpha=return in excess of benchmark is the simplistic view even professionals had 20 years ago. But nowadays you will not meet professionals in finance talking about “alpha” the way you do it. And if it happens, may I suggest you walk away from these “professionals”? 😉

  7. Thierry permalink
    September 19, 2011 11:59 am

    Damian :
    I don’t agree – alpha is the return above a benchmark. If you have an active, long-only manager selecting stocks from the S&P500 and he or she manages on beating the index, they have generated a positive alpha without using any leverage. Maybe you mean leverage in a different way?

    Sorry, forgot to answer you question… I am pathetic.

    Wrt stock picking, this is again LEVERAGED BETA. Why? because with stock picking you necessarily take on an exposure to a different beta than the benchmark (momentum, value, industry, sector, add your own…), and as you take it on top of your benchmark exposure, you end up being leveraged!

  8. September 28, 2011 8:21 am

    Lo’s paper on the first Global Alpha crisis in 2007 may provide insight. The paper provides some insight into a crowded strategy and position unwinding which had knock on effects. The paper is entitle “what happened to the quants.” and is a must read for any fund of fund allocator or risk manager or operator in a tightly defined strategy with limited carrying capacity.

    http://web.mit.edu/alo/www/Papers/august07.pdf

Trackbacks

  1. Sunday links: mock capitalism | Abnormal Returns

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: