In the corporate pension world, liabilities are discounted using a high quality corporate bond yield curve. Historically, the pitch for Liability Driven Investing (LDI) has focused on the long duration of the liabilities and the need to use extended duration products to help hedge the interest rate risk. The latest hot topic is the other piece to the discount rate which is the credit spread. Given that spreads are at significant highs, many consultants are pushing pension committees to invest in long-term corporate bonds. This shortens the duration of the portfolio but increases the exposure to corporate spreads and should help the funded status of the pension fund if and when spreads go down. Overall I do not have too much of a problem with this concept but I do take issue with how consultants are pitching it to their clients. Since most of their client’s fixed income allocations are in treasuries, consultants will go in and say that these pension funds are only hedging 5-15% of the credit risk embedded in the pension liability. Finally, I’ll get to my point. Most of these pension funds have allocations of 30-40% US Equities. I haven’t done the analysis but I would assume there is a significant correlation between corporate spreads and US Equities which would increase the “credit hedge” that these consultants are trying to quantify. Does anyone know of a good paper that may discuss the correlation between corporate spreads and US Equities? I believe this type of analysis would be interesting given that the analysis would have to remove some of the noise (i.e. interest rates) before determining a proper correlation. Any thoughts would be greatly appreciated.
Can you just pull 30 years of S&P 500 history and credit spreads for (pick your favorite rating) and do a =correl() ? Separately, you might look at the KMV work (an attempt to predict credit strength from equity price). In general though there’s so much noise and irrationality in the equity markets that you might find little useful information content there.
Interesting thought. I’m sure there’s plenty of literature in the High Yield world that talks about equity and HY performance, and presumably HY spreads and ordinary investment grade spreads are at least somewhat correlated. In terms of prices, corporates and equities should be weakly or moderately correlated (FI price movements and Equity price movements are typically 0.2 correlated). In terms of returns, Credit Spreads and Equity should be negatively correlated, because as equity performs well and gives high returns, the risk of not being paid back falls. I wonder how to reconcile these two opposing lines of thought. One thing to note is that equities tend to have a very long duration (on the order of 40 or 60), compared to fixed income duration (FI duration is usually <= years to maturity). If the fund’s liabilities have durations longer than achievable using Treasuries or Corporates (say a government pension fund), then equities may be in the mix to increase portfolio duration as well as provide inflation protection. In that case, the optimal allocation will have more equities in the mix than you would think if you only considered the potential correlations in prices.
I don’t understand your 3rd paragraph? How do they get 5-15%? are they saying treasuries are a hedge for their equitites?
Good question and good answer. My answer is more of a long winded one. But I will state it succinctly. If you structure your portfolio so that it compartmentalizes among liab-hedging, beta, and alpha you can address this question. the liab-hedging portion is in place to track liabilities including changes due to interest rates and credit speard. if there is no empirical evidence of equities having either duration then perhaps they belong in one of the other 2 buckets and not the liab-hedging bucket. having said that i think i will go ahead and do the correl analysis Darien suggested just to see what its been in the past. in my portfolio construction howvere i would keep equities out of the liab-hedge bucket.
bchadwick Wrote: ------------------------------------------------------- > Interesting thought. I’m sure there’s plenty of > literature in the High Yield world that talks > about equity and HY performance, and presumably HY > spreads and ordinary investment grade spreads are > at least somewhat correlated. In terms of prices, > corporates and equities should be weakly or > moderately correlated (FI price movements and > Equity price movements are typically 0.2 > correlated). In terms of returns, Credit Spreads > and Equity should be negatively correlated, > because as equity performs well and gives high > returns, the risk of not being paid back falls. I > wonder how to reconcile these two opposing lines > of thought. > > One thing to note is that equities tend to have a > very long duration (on the order of 40 or 60), > compared to fixed income duration (FI duration is > usually <= years to maturity). If the fund’s > liabilities have durations longer than achievable > using Treasuries or Corporates (say a government > pension fund), then equities may be in the mix to > increase portfolio duration as well as provide > inflation protection. In that case, the optimal > allocation will have more equities in the mix than > you would think if you only considered the > potential correlations in prices. bchad, i think you are using the word ‘duration’ to describe ‘maturity’. I dont think equities change by 40-60% for 1% change in interest rates (duration measure).
KJH Wrote: ------------------------------------------------------- > I don’t understand your 3rd paragraph? How do > they get 5-15%? are they saying treasuries are a > hedge for their equitites? No. These accounts have some allocation to Barclay Agg (25% Corps) and they do credit a little to US equities but very a very small amount. My issue was that the consultant is saying that the Pension fund is very poorly hedged against spreads going down and I believe this is not necessarily the case since equities will tend to increase in this environment.
needhelp Wrote: ------------------------------------------------------- > Good question and good answer. > > My answer is more of a long winded one. But I will > state it succinctly. If you structure your > portfolio so that it compartmentalizes among > liab-hedging, beta, and alpha you can address this > question. the liab-hedging portion is in place to > track liabilities including changes due to > interest rates and credit speard. if there is no > empirical evidence of equities having either > duration then perhaps they belong in one of the > other 2 buckets and not the liab-hedging bucket. > > having said that i think i will go ahead and do > the correl analysis Darien suggested just to see > what its been in the past. > > in my portfolio construction howvere i would keep > equities out of the liab-hedge bucket. Good point but I’ve never been a fan of the bucket concept. The entire portfolio is going to move with certain scenarios and I believe you have to take this into account when evaluating the portfolio against the liabilities. I have an account that has a liability duration of 20! Very long and the consultant is pushing long-term corporates which will significantly reduce the funds duration hedge. I know it is hard to believe that rates could go any lower but LDI is a strategic strategy not a timing the market strategy.
needhelp Wrote: ------------------------------------------------------- > > bchad, i think you are using the word ‘duration’ > to describe ‘maturity’. I dont think equities > change by 40-60% for 1% change in interest rates > (duration measure). I meant interest rate sensitivity, but I can see how you might think I had it confused. As a rule of thumb, I tend to remember that duration <= years to maturity because that helps me remember that a zero coupon bond has Duration=Maturity, and a coupon bond has less, how much less depends on the timing and size of coupons. But equity has a duration too… look at this (shamelessly ripped from John Hussman’s analysis, so it’s not mine): (http://www.hussmanfunds.com/wmc/wmc040223.htm) Stocks have a very long duration While the concept of duration is generally used in bond market analysis, stocks also have a duration. In the case of stocks, duration measures the percentage change in stock prices in response to a 1% change in the long-term return that stocks are priced to deliver (see Estimating the Long-Term Return on Stocks). For the stock market as a whole, the modified duration is simply the price/dividend ratio, which for the S&P 500 is currently about 62. The duration itself is about 67 (the precise figure depends on the exact return that you believe stocks are priced to deliver). [Geek’s note: to see this, consider the dividend discount model P = D/(k-g). Differentiate with respect to k to get dP/dk = -D/(k-g)^2. Divide through by price, which is D/(k-g), and then substitute P/D for 1/(k-g). Notice that this result is independent of g. For stocks that don’t pay a predictable stream of dividends, you have to calculate duration explicitly from the stream of expected free cash flows, but for blue-chip indices, the price/dividend ratio is an excellent proxy for modified duration.]
bchadwick Wrote: ------------------------------------------------------- > needhelp Wrote: > -------------------------------------------------- > ----- > > > > bchad, i think you are using the word > ‘duration’ > > to describe ‘maturity’. I dont think equities > > change by 40-60% for 1% change in interest > rates > > (duration measure). > > > I meant interest rate sensitivity, but I can see > how you might think I had it confused. As a rule > of thumb, I tend to remember that duration <= > years to maturity because that helps me remember > that a zero coupon bond has Duration=Maturity, and > a coupon bond has less, how much less depends on > the timing and size of coupons. > > But equity has a duration too… look at this > (shamelessly ripped from John Hussman’s analysis, > so it’s not mine): > > (http://www.hussmanfunds.com/wmc/wmc040223.htm) > > Stocks have a very long duration > > While the concept of duration is generally used in > bond market analysis, stocks also have a duration. > In the case of stocks, duration measures the > percentage change in stock prices in response to a > 1% change in the long-term return that stocks are > priced to deliver (see Estimating the Long-Term > Return on Stocks). For the stock market as a > whole, the modified duration is simply the > price/dividend ratio, which for the S&P 500 is > currently about 62. The duration itself is about > 67 (the precise figure depends on the exact return > that you believe stocks are priced to deliver). > > thanks bchad. i may have seen this before. but i think there is no empirical evidence to support this.
I think part of the empirical issue is that there are plenty of non-dividend paying stocks that mess things up. Also, the direct interest rate effects are probably difficult to separate from a bunch of indirect effects. I agree with you that it seems implausible that a 25 bp rise in interest rates should make stock values go down by .25*60 = 15%, so something else must be happening in the mix. Presumably stocks will also have fairly high convexity. I do think that it is worth remembering that equities can have duration, though, and that the duration could potentially be quite large.
Rates went from 5.25% to 0.25% in less than 2 years. According to that research, the market should be up 300%?