SD Co. pension responds to my story
Lee Partridge, the investment consultant for the $7.2 billion San Diego County retirement fund, responded at yesterday’s board meeting to my story that appeared April 4 in the Voice of San Diego about the fund’s $2.5 billion bet on leveraged Treasuries.
Lee and I spoke this morning and he sent along the written response that he presented to the board, which appears below. You can also watch his presentation at yesterday’s board meeting by clicking here.
From: Lee Partridge
Sent: Thursday, April 08, 2010 12:41 AM
To: Brian White
Cc: Collins, Jeff_CoSD; Lisa Needle
Subject: RE: Fwd: voice story
Brian,
I read through the Voice of San Diego article again and have a few comments. I’ve also copied Jeff for your convenience (hi Jeff). There are a couple of misleading or false statements but generally I thought most of what Seth said was true. Nonetheless, he’s overemphasized the negatives. My former boss used to say that if there’s a one percent chance that something could go wrong and you spend 50% of a presentation talking about it, you haven’t represented the truth accurately, you’ve distorted it. I think Seth’s point was that if Treasuries go down a lot in price, we’ll lose money on them. That’s tautological and not very helpful.The reality is that Treasuries will likely fall only if the rest of the portfolio is doing well. Treasuries have very stable, negative correlations to the risky assets in the portfolio and they exhibit relatively low price volatility. We’re trying to increase the price volatility of Treasuries sufficiently to offset losses that we’ll take on stocks, credit, private equity, emerging markets, real estate and natural resources in a bear market. My fear is that we still don’t have enough Treasuries to protect the downside of the plan—not that we have too many. Investors are generally way too complacent about risky assets. This is not a risk seeking strategy or a speculative bet, it’s a way to control downside risk that emanates from economic deterioration.
No matter what my asset allocation is it would be easy to paint a bleak picture if a reporter said something like, “Partridge says that if market conditions force each of these markets to fall the fund could lose billions of retiree dollars forcing taxpayers to foot the bill.” Technically that would be true there is no true information contained in the statement. If we want to be certain that we won’t lose money than we either need to increase contributions dramatically or lower benefits. This is less risky than the portfolio you had.
Taking into account the whole balance sheet it’s even a stronger argument. I took the benefit cash flows that Paul Angelo sent me and discounted them back at various discount rates. Those discount rates are pretty good proxies for the relative riskiness of the asset allocation strategy. I believe the red line is what you generally focus on the most. Note that the liabilities move from around $9 billion at an 8.25% discount rate to approximately $17 billion if discounted at 4%. That is the risk free rate of return with which we could match our liabilities with perfect certainty. It’s what Paul refers to as the settlement value of the liabilities.
The settlement value of the liabilities rises and falls by about 15% for every 1% change in interest rates. Using the fair value of our liabilities at a 7% assumed rate of return that would translate to a $1.6 billion market movement for every 1% change in interest rates. Also, I believe part of the pension was funded with fixed rate pension obligation bonds, which further exacerbates this picture. To place the Treasury futures position in context. The value of the Treasury futures position will rise and fall by approximately $200 million for every 1% change in interest rates. So, in addition to offsetting the risky assets, the Treasury position also offsets approximately 13% of the interest rate risk of the liabilities (not including any pension obligation bonds that were issued to fund the plan). The county and the plan has already taken a huge bet on interest rates but it’s not through the Treasuries it’s in the overall mismatch of assets and liabilities. We can save that for another discussion but it’s very important to keep in mind since this helps offset some of the problem embedded in the plan.
This strategy is about fundamentally diversifying the portfolio so that it does well during periods of both inflation and deflation, high growth and low growth, and during periods of still waters or crisis. This strategy isn’t rolling the dice, everyone else is rolling the dice and we look different because we’re being prudent. Why would 65% in public equities make sense to anyone managing public money? 2008 proved how painful that exposure could be. This strategy hardly ever underperforms and when it does it’s during periods of rapid equity growth that is generally accompanied by inflation. Again, I go back to the question, would you be willing to underperform your peers in the one scenario where you will be most likely to meet the plan obligations with no problem if you could outperform your peers in almost any other scenario?
I’ve taken a few of the comments from the article and commented on them.
(Click image below to enlarge:)
These comments elaborate and support the idea of risk parity in general. I basically favor the idea for SDCERA and I think it is well suited for SDCERA and other public funds. I hope these remarks helps others to understand the concept and that is my intent.
“Risk parity” was popularized by Ray Dalio of Bridgewater Associates. The idea is to lever up bond (or Treasury) exposure to a risk level approximating the riskiness of stocks. This increase in Treasury exposure provides some downside protection for an investment fund, especially protection against recession. The appearance of heavy duty fear in the market will buoy Treasury prices and exposure (through derivatives) will have gains. Such insurance, like all insurance, is not free: If it were free, then all portfolios would be so insured and hedged completely. Even so, risk parity, expressed in Treasuries, enhances a portfolio.
There are other ways of protecting a fund against the downside risk and recession. One could use the $250 million required to secure the SDCERA’s implementation to buy put options on equity which choice has the benefit of limiting costs explicitly. Another method – and my favorite – would be to develop flexible trading rules which allow adjustments to asset class allocations at the margin. This approach can be very productive net of costs.
Another approach would simply diminish the equity allocation in the unleveraged portfolio relative to bonds, resulting in lower overall volatility but nonetheless high efficiency. The problem is that this approach would not provide the targeted or required level of return and this is the rub that risk parity strategies address.
Risk parity requires that the fund have capital invested in excess of the assets of the portfolio. The program recommended at SDCERA is for capital exposure of 135% meaning that the fund will have $1.35 invested for every $1.00 of assets. The excess exposure, however, is expressed as Treasuries which are considered risk free, not invested in risky stocks, for example. Stretching beyond the assets of the plan in this manner allows a fund, however, to retain its exposure to risky assets and to expect the rewards of such investing rather than delimiting such exposure in the instance in which leverage is not allowed. Leveraging Treasuries allows for more risky investing than would otherwise be possible or allowable, depending on the fund. The Treasury exposure, in effect, shelters risky investing. The addition of leveraged assets, however, means more absolute return to the fund because Treasuries earn return.
Both Partridge and the consultant say that the program is aimed at “controlling” risk, not increasing risk, yet all investment pros know that any increased return means increased risk. What account can be offered for this contradiction? The answer is that Partridge and the consultant use the term “risk” in a sense that is different from the formal definition of risk in finance. They do not mean that volatility is dampened because it is not. They mean that bad outcomes on the downside are tempered by owning excessive Treasuries. This is true but it is not the same as saying that the program diminishes risk. Adding Treasries beyond 100% of assets does not reduce portfolio volatility relative to the 100% unleveraged portfolio. But, it does offset a portion of losses expected on the downside, especially a downside based upon fear of recession. Of course the offset of stable or increasing Treasury prices is not a free good but the cost depends on interest rates which are currently low.
While risk parity does not require explicit borrowing of cash for investment from a bank, the strategy involves implicit borrowing to the extent that the futures exposure is uncollateralized by cash assets. Partridge and the consultant and staff know very well that such positions incur implicit borrowing costs, borrowing of cash as it were from the exchange, but probably thought it best not to try to explain this and I sympathize. Suffice it to say that one cannot have exposure of 135% without borrowing, by definition. Excess capital exposure, like excess capital, is not free.
The academically interesting aspect of the risk parity notion, in my view, is that such a strategy of leveraging an asset class necessarily changes the risk/return tradeoff of the 100% unleveraged portfolio to a lower, less efficient level. The portfolio mix is changed to something suboptimal assuming the unleveraged mix was selected because it was optimal or nearly so. This is necessarily true because the relevant attribute of leverage for optimization purposes is the cost of funds, not the leverage multiple which is not an input for optimization. The cost of funds is a drag on performance regardless of upside or downside, that is, a cost is a cost in any case. There are no two ways about this. In order for the portfolio to perform better on the downside in particular, something must be given up in upside potential. This is simply how optionality, and hedging, works. Even Wikipedia mentions the implicit borrowing of derivative transactions.
The choice of a strategy for protecting returns on the downside of the market is a matter of taste and style. The choices usually made reflect the relative flexibility of the mechanics for managing the portfolio. I mean that, while it is obviously better to have downside protection only when you need it, and not to have it when you don’t, most funds are not set up to change exposures on a day to day basis in this manner. As a consequence, a perfectly good hedge during a downside, just comes along for the ride on the upside. The risk parity strategy is not highly flexible, i.e. it is a permanent feature of the portfolio.
While not flexible, and not perfect, such an implementation is a whole lot better than taking losses hand over fist in another credit debacle. While I don’t think another credit debacle is economically plausible going forward, few are prepared to bet against it at this point. While someone might object that the strategy is, in effect, fighting the last war, the last war was so horrendous that it cannot be overlooked. More likely, and more troubling for a risk parity strategy in Treasuries, is the possibility of rising rates which could negate the downside protection strategy very quickly, incurring losses rather than offsets.
Risk parity was implemented in the SDCERA portfolio through the All Weather Portfolio investment and the idea was presented to the board on several occasions in the past. I think it is a good idea for downside risk control but it is not magical and not free. 100% does not become 135% willy nilly.
Banks and S&L’s are critically concerned with asset/liability matching for reasons related to regulations, risk capital and net interest income, the business of banking, none of which much concerns a public pension. I have never been convinced that asset/liability matching, as such, is very important for a public fund. In my own view, successful active management of assets is the best hedge for liabilities.
Just wanted to correct some misstatements in my prior comment. It appears that I am talking to myself, in any case.
The levered Treasuries adds a hedge and dilutes the weights of “risky” assets in the portfolio. The overall levered portfolio may be efficient and optimal but only with respect to a lower risk tolerance. It is a hedge. Without the leverage, the portfolio is more aggressive and has a higher level of expected return with a higher livel of risk. As with any hedge, you take from the upside potential to offset the downside risk. The expected return on this portfolio is lower both by virtue of the risk and the embedded costs of hedging.
Anything I’ve written previously that suggests something different from these summary remarks, is wrong or not properly construed. In any case, this hedge is a good idea and something we were looking at SDCERA long ago: We opted for dynamic rebalancing, small changes in allocations, which would help on the downside (as it did) and help on the upside (which it did). A hedge of TReasuries are one way and only offset the downside while incurring shortfalls in return on the upside.
When interest rates take off as they surely will at some point (the alternative would be Japan — ten years of deflationary recession!), the SDCERA will have to fund losses on the Treasury futures with cold cash and this is very onerous, as I know so well.
DD