When you expect curve to flatten in the near future, do you want to increase convexity? Barbell portfolio has more convexity than laddered point portfolio or single asset portfolio. Flatten curve helps increase asset value of barbell portfolio more relative to liability than other asset portfolios vs liability. Therefore, I believe if expectation is flatten curve, then adding convexity would be the way to go. Not 100% certain my reasoning is correct.
If you believe the pure expectations theory for yield curves, then a flat yield curve suggests that interest rates are not going to change in the near future. In that case, convexity won’t help you, so you should sell it.
Kaplan book states yield curve flattening generally is good. Long duration assets appreciate more than long duration liabilities. Short duration assets decrease more than short duration liabilities. Overall, since long duration assets tend to have stronger impact on the portfolio, near future yield curve flattening is considered to be generally good. I still believe it would be better to buy convexity instead of sell convexity with near future yield curve flattening, because yield curve flattening benefits barbell portfolios, which has very high convexity. Anyone agrees?
You are correct. To minimize structural risk (dispersion), you must sell convexity in my example.
If you EXPECT the yield curve to flatten that means you expect rates to change. Barbell outperforms when the curve flattens. You increase convexity.
Think, if you own a 2-30 portfolio and long term rates fall (flattening) you’re 30 year bond will appreciate significantly.
My mistake for misreading the original question.
If you expect rates to change, you want to buy (positive) convexity. If you expect rates to remain unchanged, you want to sell (positive) convexity.
Is it true that if you expect rates to unchange or immunizing a single liability, then you should minimize portfolio convexity?
You should maximize convexity in situations of large parallel shifts, steepening twist, flattening twist, positive butterfly twist, and negative butterfly twist? I actually think you should sell convexity with steepening twist or positive butterfly twist, because barbell portfolio will become worst than laddered or bullet portfolios, because barbell portfolio has most convexity. Not very sure though.
A “flat yield curve” and an “expectation of a flattening yield curve” are 2 different thing. Sell Convexity in a flat or expected to remain flat yield curve environment but BUY convexity if the yield curve is expected to flattening in near future.
This issue has come up a lot. You want your convexity to be higher than that of the liabilities that you are immunizing but minimally larger. Maximizing convexity in general creates structural risk of the immunizing portfolio and therefore you wouldn’t want a portfolio simply with the highest convexity but rather the one with slightly higher convexity than the liability.
interesting to see the discussion on here. Your reasoning appears correct if you are constrained to keep duration the same or to match liability duration. You didn’t say that but seem to be assuming it (it is a common assumption so I can see why)
if you are not duration constrained you’d want to go to a longer duration portfolio (and assuming you had the courage of your convictions - bigger potential reward but bigger potential risk if you are wrong). Or look at key rate durations and invest where the biggest drops are expected but generally flattening means long rates are coming down more (relative basis)
i’ve seen practice questions that get at both ideas so you probably want to understand both if you are studying.
In the likelihood of increase uncentanty of yield curve or interest rate movement. One would like to buy convexity to protect against the increase volatility