BENN EIFERT: There was really a large psychological overhang after the credit crisis- psychological and real, where people remembered how bad it was to be caught short convexity during a huge equity market selloff.

There was academic research obviously on the volatility risk premium and the historical tendency of implied volatility to be above subsequent realized volatility on average.

I think it's a very long cycle, though, because the types of folks that have large state pensions. From the time where they get interested that this might be a good idea to the time that they actually are able to decide what to do, approve the program and roll it out can be years.

MIKE GREEN: Mike Green. I'm here in Las Vegas at the EQDerivatives Conference. I'm going to sit down with Benn Eifert of QVR Advisors. Benn is a PhD, absolutely brilliant in the option space. And we're going to spend a lot of time talking about some of the technical features of option rebalancing and the structural vol sellers in the marketplace. In addition, we're going to try to talk about some of the characters that are involved, hopefully, add a little bit of color to it. Looking forward to sitting down with Benn, I hope you enjoy it.

Mike Green. I'm here in Las Vegas at the EQDerivatives Conference with Benn Eifert of QVR Advisors. Sensor.

BENN EIFERT: Quantitative volatility research.

MIKE GREEN: Alright, and then is a volatility specialist. I play one on TV occasionally, but you do the real deal. So, we just got off of a panel where the two of us presented and you brought up one of my favorite topics, this idea of volatility suppression and volatility selling. Help me understand how you got to a role in life where you focus on this and what it means to you and the day to day dynamics of your activities.

BENN EIFERT: Yeah, absolutely. So, my background is I have a PhD in Economics from Berkeley. I was an old school, emerging markets macroeconomist back in the day. I ended up on the Wells Fargo prop desk for a while, which was quite fun. And nobody actually knew that there was a Wells Fargo prop desk, which makes it even more fun. And then ended up running macro volatility strategies there and elsewhere and started this firm a couple of years ago out in San Francisco.

MIKE GREEN: What period were you running the macro volatility strategies at Wells Fargo?

BENN EIFERT: That was 2010, '11, '12, '13 at Wells Fargo and then a partner and I left Wells Fargo to start a fund in New York called Mariner Courier, a Mariner Investment Group platform managing a global volatility portfolio for three years, and then I moved back to California to start my own firm.

MIKE GREEN: Now, when you moved on to the prop desk, the macro strategist, you were working on the prop desk prior to the GFC? Is this-

BENN EIFERT: We're right in the middle of the GFC was when I first was getting involved.

MIKE GREEN: So, you came out of academia, you went into chaos?

BENN EIFERT: Exactly. It was great fun. And actually, the Wells Fargo prop desk was a really interesting place. Because you would think of it as probably the 23rd best prop desk on Wall Street at the time, but it also happens-

MIKE GREEN: It's officially 22nd.

BENN EIFERT: Maybe 22nd. Yeah, exactly. But that's right, Lehman got taken out of the mix. But importantly, it was the prop desk at the bank that didn't know what a synthetic credit derivative was. And had no- obviously, had real estate exposure and home loan exposure but had none of the really toxic stuff. And so, the Wells Fargo generally was fine, actually, in late 2008. And the bank was willing to really support the desk to go take advantage of a lot of the opportunities that were out there in the market in late '08, early '09. And so, that was great. Whereas many of the higher ranked prop desks in the world were getting liquidated and blown out of their positions because their banks were in trouble.

MIKE GREEN: Well, and there was a substantive risk change then as well as in terms of the regulatory environment. So, while you were able to take advantage of some of this because you had a relatively clean book, you wouldn't have been able to engage in the type of behaviors that existed pre-2007.

BENN EIFERT: Yeah, that's exactly right.

MIKE GREEN: So, when you came in, and you talked about taking advantage of these opportunities, what were those opportunities? What were you seeing?

BENN EIFERT: Well, I think the Wells Fargo prop desk was the last buyer standing for convertible bonds and for secured loans. And for a lot of the credit products trading at five or 10 cents on the dollar at the bottom of the market. And think of how wide for example, cash synthetic basis got to the credit markets though, because of funding. That's really a funding trade. We look at bonds trading a much wider spreads than the CDS that hedges those bonds, because effectively, if you're going to buy a bond and buy CDS against it, you're taking counterparty risk and you have to post balance sheet and you're getting paid a tremendous amount of money if you have any balance sheet.

And Wells Fargo was one of the few folks that actually had balance sheet to deploy. So, a lot of those types of trades, particularly on the credit side were huge opportunities. Within equity land, you can also think of, if anything about the dividend swap market, there was a huge basis between index dividend pricing and the pricing of all the single name dividends that aggregated up to be equivalent to the index dividends and risk. And that came because there was one very, very large player that was long way too much of the index and had to sell it in a huge fire sale, and there was nobody to buy it because everybody again was getting shut down.

So, many of those types of opportunities where if you have balance sheet and risk-taking capability, you could take great advantage of the situation, because so few people have balanced sheet and risk-taking capability at that point.

MIKE GREEN: That was exactly my experience. And I was extraordinarily fortunate and being in a situation where I also had some access to balance sheet, although, for reasons similar to you describing, I was forced to basically create an overlay on top of that. The exact same dynamics that blow out of basis, that effectively meant that people didn't have the capacity to hold what they otherwise knew were good investments at normalized by and large circuit 2012 into 2013. How do you think about the dynamics of the structural cleaning of that market, the cleaning of balance sheets as the passage of time, and the shift that we've seen post-2012, 2013 on really, in terms of how people have chosen to trade these vehicles?

BENN EIFERT: I think over time, what you've seen, you particularly coming to the point you asked about systematic option trading, there was really a large psychological overhang after the credit crisis- psychological and real, where people remembered how bad it was to be cut short convexity during a huge equity market selloff, or even folks who weren't short convexity, but just were long a bunch of stock and it went down 60. It was incredibly painful. And it created really a persistent bid for options by hedge funds, by real money institutions, by pension funds.

You saw the rise of tail hedging programs, systematic tail hedging programs and tail funds, where investors wanted to be protected against the next leg down and everybody was thinking about what's the next thing that's going to take the market back down 50%. And that, not surprisingly, led to a lot of that type of risk being very expensive, because there was this one-way market for protecting your tails across the board. And that, to your point, really started to shift by, say, 2012, 2013, when a lot of those institutional programs had maybe budgeted one or 2% per year for loss against those programs, I actually lost four or five. And their equities were doing well.

So, it was okay. But after several years of that, just the patience for hedging fell apart. Too much money have been lit on fire. And one by one, those big tail risk hedging programs got shut down. And they were replaced by the opposite in some ways, by option selling strategies among large institutions, not necessarily tail risk selling but think of index call overwriting, for example, which was something that investors have participated into some extent or another for quite a long time. It was always popular to say you own a stock and you're willing to sell off the upside beyond some point and retail investors like to do things like this.

But institution large pension funds started, more and more being in overwriting say 10% or 20% of their equity portfolios starting in about 2012, 2013. Selling cash secured puts against their equity portfolios. And those programs have grown and grown and grown very steadily, say doubling every two years for the last six years or so. And now, I think you really look at some supply and demand in the option marketplace and it's completely flipped from the post-crisis hangover, where there's a very large amount, arguably too large for the capacity of the market to support it of short term, relatively near the money option selling by large institutional investors as an overlay within their portfolio.

MIKE GREEN: This is one of these interesting switches that occurs. So, if we come out of 2009, and we look at what I would, by and large, call the Black Swan market. The crisis hedging opportunities in that market on a pure size basis grew to about $60 billion worth of allocated capital that was dedicated to buying, all hell breaking loose puts, whether it was in fixed income, or it was an equity, it didn't really matter. So, there's this tremendous bid to actually buy that convexity and to buy that protection.

And most people tend to think about options as either buying protection or as a form of buying leverage to the tops. I think the stock is going to do great. And therefore, I buy a call option because I want to make a lot more money. But what you're highlighting is, is that the vast majority of the institutional exposures have now shifted to using options to generate what we would probably call yield enhancement. Why are they doing this?

BENN EIFERT: I think it's been a long process of evolution. There was academic research, obviously, on the volatility risk premium and the historical tendency of implied volatility to be above subsequent realized volatility on average. Has been around for a while. You started to see some of the pension fund consultants writing white papers based on that evidence, say even in the mid to late 2000s, early 2012, 2013. And the process of those consultants working with pension fund clients and educating them on what they saw as this opportunity took a period of time.

And then, as with anything, large institutional allocators move slowly but once they start moving, and something becomes accepted and popular and consensus so that as a pension fund allocator, your boss isn't going to say what the heck are you doing over there? That's some very unusual thing. Now, it's actually very common place and you wouldn't be questioned. So, the answer you had, this persistent growth.

If you were to look at a long term back test, and AQR has some nice papers on this. If you're going to take a 30-year perspective and say, is there a volatility risk premium? The answer is you get is, of course, absolutely. Where is it the greatest, if you look across the volatility surface, short term options, long term options, deep tail options? The long term evidence on average will say, well, it's highest for relatively short term options, let's call it one-month, and it's highest for around the money, slightly downside puts snapped too far downside. And, like anything, you can see a historical pattern for 30 years, that's true on average.

But then if there's a dramatic amount of behavior change in response to those prices, then that risk premium can change very dramatically. And I think that that's what we've seen over the last several years, if you look at the evidence. And you need statistical models, and you need to do the work. My view is there still a volatility risk premium at the short end of the curve, but maybe it's half of vol point. So, on average, if you sell 10, you expect a realized vol of 9.5 or something like that, maybe it's three quarters of the vol point. Whereas historically, over long periods of time, it was closer to say three vol points.

So, you've cut that risk premium, the large inflows into that, into very benchmark like strategies all selling the same thing, typically one-month range, typically within 25 delta, calling to put legs relatively close to the money has changed risk premium by an order of three or four times, so cut it by 70%. And those flows continue to grow at about the same pace.

When you raise the question with pension funds or pension fund consultants, you say, have you thought about the size of the market, the capacity of the market to absorb these flows? At what point risk premium start to change? Have you done some work on whether you see risk premia changing typically? The typical responses- the folks just haven't really thought about it that way. They think of this as a risk premium that's behavioral and will always be there and should be roughly constant over time. But I think that we just know that that isn't the case. That's not how markets work, when large amounts of capital move into doing something very, very similar and very visibly changed the prices.

So, if you look at the typical shape of the S&P, just at the money volatility term structure, look at a quiet volatility regime prior to the credit crisis, you might see something like at the money one-month implied vol of 13 and the one or two-year at 15, before the credit crisis in a quiet period. Now, during quiet periods, it's eight or nine at the one- month point and maybe 17 in the back. So, these flows have dramatically steepened the volatility term structure and make the front much, much lower. So, we're changing risk premium, we're changing prices. But I think the people involved aren't really asking that question.

MIKE GREEN: They aren't asking that question. That's one component of it, too. And this is something you've heard me talk about is, functionally, all of these assumptions are predicated on the work of Mark Watson Sharpe. And their ideas are very straightforward that we can model these is if they're games of chance, they're stochastic in nature. The frequency in the past is the same as the frequency in the future. And nobody stops to say, well, what's the capacity when you talk about these models? Nobody stops to say, what's the capacity of a blackjack table or a roulette wheel? Functionally unchanged by the players that are playing the game. But those are the tools that were built. So, stochastic model, we use Monte Carlo simulations, precisely because we think that we're simulating games of chance. But that's not really what happens.

BENN EIFERT: Now, That's exactly right.

MIKE GREEN: Yeah. So, explain to me what you're doing to think differently about these risks. How do you think about things like crowding? How do you think about changing risk premiums?

BENN EIFERT: Yeah, and I think you've hit on it exactly right. The key to why finance and investing is interesting and hard is that the world is a non-stationary place. The distributions of investment returns that we're drawing from today are different than they were in 2013 and '09 and '07. Otherwise, we would really have quite a lot of data from a historical perspective, and actually could believe that we very confidently know the risk reward properties of all kinds of trades and build portfolios in practice. Actually, historical data can guide you to some extent. But you have to think you have to, in some sense, develop your structural models of what the underlying factors are that generate these outputs, and have ways to measure those over time and understanding what's changing and how much and in what way.

And so, one thing, for example, that we do is try to measure these flows in terms of what is the net buying or selling of short term volatility by customers from option market makers over time. That data doesn't literally exist, what you have to do is build a probabilistic model from trades that you can observe, what we do is we look at awareness, look at every trading, all every one of 100 million option trades that have happened over the last 20 years and assess based on where it was priced relative to where the best bid and best offer was across the theoretical volatility surface. What's the probability you think that this was a trade? That was a customer selling to a market maker, a customer buying from a market maker, or something else.

MIKE GREEN: And just as a simple illustration of that, so if you saw a trade that occurred closer to the bid, then you would presume that that was actually a situation in which they were actually offering it or trying to sell it? And the reverse is true, if it was suddenly asked?

BENN EIFERT: Yeah, that's exactly right. And of course, sometimes people are able to buy it mid-market and so forth. Typically, it'll be more sophisticated market participants who are dynamic hedgers themselves, who are working orders algorithmically, and so forth. But when a typical large overwriting firm comes to the market to execute a roll transaction, they'll quote an open outcry, 10,000 call spreads in the market and to get their size done, that won't trade at mid, that will trade down towards the good side. Or if it's big enough, and markets are a bit illiquid, maybe below what you would have seen as the best bid.

And you can identify, at least with a reasonable probability that that's, in fact, an option being sold. And you can track that over time and build up the full evidence. And what you see is roughly the pattern that we described where pre-crisis, you started to build up significant net volatility selling exposures. The crisis really squeeze that as it forced largely hedge funds and hotter money that was short volatility to cover, left a big hangover, where investors were net buyers of options for several years, that sign probably flipped to negative somewhere around 2013. And then started to go more and