You Think Hard, You Lose Clarity
Again I'm going to write a bit about macroeconomics. I want to focus on an area I poorly understand. At some point I will have to read some of the literature that is out there, but I will put that day off as long as I can.
The thing I am thinking about is my earlier distinction between deferring consumption and assuming risk. The logic is that when I buy a financial asset for $x, I am generally doing both: I am (A) spending $x less than I otherwise could have, in the hopes that I will be able to spend more in the future, and (B) taking on the risk that the asset will decline in value. (Let's focus on traditional securities and similar products, not, like, triple-short ETFs or whatever.) (And as a side note, the risk-bearing concept shows up in people's description of their investment activity when they say they want "exposure" to a particular region or sector. What they mean is that they want to bear the downside risk of losses in exchange for the upside "risk" of gains.)
Now I think these two concepts (deferring consumption and assuming risk) are fairly easy to distinguish as a conceptual matter. Of course most financial products blend the two, though you can imagine exceptions. (For instance, if my broker allows me 100% margin, then I can "gain exposure" to a security without foregoing any current consumption. In the U.S., this is generally not permitted for equity securities. Or I can sell options and use the premiums to increase my current consumption - again assuming the margin requirements are low enough that I can extract the cash.)
But while I think the concepts are reasonably distinct, I wonder if we can really separate them in practice. Here's what I mean. Imagine that I can buy either of two Treasury securities. One of them is a 6-month bill, and the other is a 5-year note. (For our purposes, the only difference between a bill and a note is time to maturity. In real life I think bills are sold on a discount basis while notes bear a coupon. Ignore this complication.) Imagine that these are TIPS (Treasury inflation-protected securities), meaning that the payment on the security is adjusted to reflect inflation. Assume for simplicity that the securities bear zero credit risk and that they reflect a completely accurate measure of inflation (so that there is truly no inflation risk with these securities). And finally, assume that the securities can't be sold or pledged. Once I've bought them, I'm stuck with them until maturity, no matter what.
Now, there are still risks with these securities. Most notably, there is interest rate risk - if interest rates go up after I buy the securities, then I will incur an opportunity cost - if I had held my money in a checking account, I could invest in the same securities at a higher interest rate. And maybe I will be compensated for that risk (more so with the 5-year note than with the 6-month bill, because the risk is greater the longer I have to wait to get my money back). But it seems as though these securities are as close to pure consumption-deferral as you can get (you are trading away $x in 2014 dollars in exchange for $y in 2014 dollars in the future), and again the compensation should be higher for deferring my consumption for 5 years than for 6 months. (Strictly speaking, the point here is that I could either defer consumption for 5 years, or defer it for 6 months with an option to extend for another 6 months and so forth. Since the rolling 6-month option gives me more opportunities to consume in the relatively near future, it should bear lower compensation, that is, a lower yield.) So in ordinary conditions we would expect the interest rate to be higher on the 5-year note than on the 6-month bill. (In rare situations the "yield curve" may be "inverted" - this just means that long-term securities actually bear a lower interest rate than short-term securities. This might happen if short-term interest rates are currently high but are predicted to drop in the future.)
Okay, very well. But now imagine the same securities, except they can be sold or pledged as collateral. And imagine there is a reasonably liquid market for the securities, meaning that in practice I can sell them quickly for a reasonable price. Now it's easy to see that I'm still bearing interest rate risk (in a sense, it's easier to see because if interest rates go up I might actually sell the securities for a loss, whereas in the previous example I was compelled to hold them to maturity).
But the crucial insight (I think!) is that I'm no longer deferring my consumption differently with the two securities. I can sell either one of them on short notice. I'm still taking on more risk with the longer-dated security, but I'm no longer really committing to wait 5 years before I can spend the money I'm investing. I could sell the 5-year note next week and buy some kidney beans or whatever. So my consumption-deferral is now roughly equal as between the two securities (although, to repeat, I am exposed to more risk, both upside and downside, on the longer-dated security).
The point here is that consumption-deferral seems to be fungible with risk-taking as an economic matter. (Extending the amount of time I am willing to defer my consumption is fungible with bearing additional interest-rate risk.) So distinguishing between the two might not be so easy to do in the real world. I ought to be able to enter into some kind of derivative contract with no money down (if my counterparty is willing to take my credit risk) that exposes me to roughly the same risk as holding these securities, but doesn't require me to defer consumption at all.
Now I'm not sure what the implications of this are. Maybe nothing? But it seems odd that consumption-deferring and risk-bearing might be fungible in terms of how they contribute to the economy and how they are compensated. (Again, I acknowledge that most securities involve both activities, so I wouldn't be surprised if it were difficult to sort them out if we tried to measure them. But I am surprised that they may turn out to be truly interchangeable as an economic matter.)
In my earlier post I wrote that in our current economic climate, risk-bearing should still be rewarded but consumption-deferral should not. (This is because we currently have a shortage of consumption, so deferring it does the economy more harm than good.) But if I'm right now, then I was wrong then. A more accurate way to put it is that certain kinds of risk-bearing should be rewarded and certain other kinds (the ones that replicate the economic features of consumption-deferral) should not. Maybe this is just another way of saying that the yield curve should be flat. Which is another way of saying that short-term interest rates should be low for a long time. Which makes sense. But it lacks the clarity of my earlier framework.
The thing I am thinking about is my earlier distinction between deferring consumption and assuming risk. The logic is that when I buy a financial asset for $x, I am generally doing both: I am (A) spending $x less than I otherwise could have, in the hopes that I will be able to spend more in the future, and (B) taking on the risk that the asset will decline in value. (Let's focus on traditional securities and similar products, not, like, triple-short ETFs or whatever.) (And as a side note, the risk-bearing concept shows up in people's description of their investment activity when they say they want "exposure" to a particular region or sector. What they mean is that they want to bear the downside risk of losses in exchange for the upside "risk" of gains.)
Now I think these two concepts (deferring consumption and assuming risk) are fairly easy to distinguish as a conceptual matter. Of course most financial products blend the two, though you can imagine exceptions. (For instance, if my broker allows me 100% margin, then I can "gain exposure" to a security without foregoing any current consumption. In the U.S., this is generally not permitted for equity securities. Or I can sell options and use the premiums to increase my current consumption - again assuming the margin requirements are low enough that I can extract the cash.)
But while I think the concepts are reasonably distinct, I wonder if we can really separate them in practice. Here's what I mean. Imagine that I can buy either of two Treasury securities. One of them is a 6-month bill, and the other is a 5-year note. (For our purposes, the only difference between a bill and a note is time to maturity. In real life I think bills are sold on a discount basis while notes bear a coupon. Ignore this complication.) Imagine that these are TIPS (Treasury inflation-protected securities), meaning that the payment on the security is adjusted to reflect inflation. Assume for simplicity that the securities bear zero credit risk and that they reflect a completely accurate measure of inflation (so that there is truly no inflation risk with these securities). And finally, assume that the securities can't be sold or pledged. Once I've bought them, I'm stuck with them until maturity, no matter what.
Now, there are still risks with these securities. Most notably, there is interest rate risk - if interest rates go up after I buy the securities, then I will incur an opportunity cost - if I had held my money in a checking account, I could invest in the same securities at a higher interest rate. And maybe I will be compensated for that risk (more so with the 5-year note than with the 6-month bill, because the risk is greater the longer I have to wait to get my money back). But it seems as though these securities are as close to pure consumption-deferral as you can get (you are trading away $x in 2014 dollars in exchange for $y in 2014 dollars in the future), and again the compensation should be higher for deferring my consumption for 5 years than for 6 months. (Strictly speaking, the point here is that I could either defer consumption for 5 years, or defer it for 6 months with an option to extend for another 6 months and so forth. Since the rolling 6-month option gives me more opportunities to consume in the relatively near future, it should bear lower compensation, that is, a lower yield.) So in ordinary conditions we would expect the interest rate to be higher on the 5-year note than on the 6-month bill. (In rare situations the "yield curve" may be "inverted" - this just means that long-term securities actually bear a lower interest rate than short-term securities. This might happen if short-term interest rates are currently high but are predicted to drop in the future.)
Okay, very well. But now imagine the same securities, except they can be sold or pledged as collateral. And imagine there is a reasonably liquid market for the securities, meaning that in practice I can sell them quickly for a reasonable price. Now it's easy to see that I'm still bearing interest rate risk (in a sense, it's easier to see because if interest rates go up I might actually sell the securities for a loss, whereas in the previous example I was compelled to hold them to maturity).
But the crucial insight (I think!) is that I'm no longer deferring my consumption differently with the two securities. I can sell either one of them on short notice. I'm still taking on more risk with the longer-dated security, but I'm no longer really committing to wait 5 years before I can spend the money I'm investing. I could sell the 5-year note next week and buy some kidney beans or whatever. So my consumption-deferral is now roughly equal as between the two securities (although, to repeat, I am exposed to more risk, both upside and downside, on the longer-dated security).
The point here is that consumption-deferral seems to be fungible with risk-taking as an economic matter. (Extending the amount of time I am willing to defer my consumption is fungible with bearing additional interest-rate risk.) So distinguishing between the two might not be so easy to do in the real world. I ought to be able to enter into some kind of derivative contract with no money down (if my counterparty is willing to take my credit risk) that exposes me to roughly the same risk as holding these securities, but doesn't require me to defer consumption at all.
Now I'm not sure what the implications of this are. Maybe nothing? But it seems odd that consumption-deferring and risk-bearing might be fungible in terms of how they contribute to the economy and how they are compensated. (Again, I acknowledge that most securities involve both activities, so I wouldn't be surprised if it were difficult to sort them out if we tried to measure them. But I am surprised that they may turn out to be truly interchangeable as an economic matter.)
In my earlier post I wrote that in our current economic climate, risk-bearing should still be rewarded but consumption-deferral should not. (This is because we currently have a shortage of consumption, so deferring it does the economy more harm than good.) But if I'm right now, then I was wrong then. A more accurate way to put it is that certain kinds of risk-bearing should be rewarded and certain other kinds (the ones that replicate the economic features of consumption-deferral) should not. Maybe this is just another way of saying that the yield curve should be flat. Which is another way of saying that short-term interest rates should be low for a long time. Which makes sense. But it lacks the clarity of my earlier framework.
posted by James | 7:19 AM
2 Comments:
I'm probably missing something but have you shown both directions of the equivalence? One can still imagine pure risk-taking activities that do not involve deferring consumption. For instance you could sell a kidney for cash for beer; or, more in your framework, you could be paid a risk premium upfront: as the expectation values of the risky and the nonrisky activities are the same, you are not deferring consumption in any obvious sense... It does not seem that encouraging this sort of risk-absorbing necessarily involves encouraging deferred consumption.
I suppose one diagnosis of the recent past is that a lot of actors, esp. individuals, have wanted to defer consumption without assuming risk. However the socially useful ways of deferring one's consumption -- by investing in potentially valuable new enterprises -- are all high-risk. The natural way to resolve this tension is to have risk-averse investors pool into enormous clumps that can use diversification and the law of large numbers to mitigate risk. ("Diversification" etc. Of course for the market to function as an information-conveying mechanism you also need some people to gamble with some of their money on the basis of "insider" information.) For some reason this economy-of-scale approach is not generally adopted. One particularly socially worthless/harmful way in which people defer consumption and assume risk is by buying a house.
A few points.
1. I agree that the equivalence doesn't run both ways as a general matter. My point is just that the secondary market for financial products creates a situation in which consumption-deferral becomes economically indistinguishable from a certain kind of risk-taking, and so you can (if you choose - and maybe you don't have a choice) analyze the whole thing in terms of risk.
2. What you are describing is the banking system, as traditionally conceived. More recently the formal banking system has come to be seen as a subset of the sector of the economy that borrows short and lends long (with a certain amount of activity being conducted by "shadow banks"). Margin lending is also a bit like what you describe, in the sense that the borrower is trying to profit from what you call "insider information" while the lender is simply trying to earn a short-term rate of return in a relatively safe manner (because margin loans are typically highly over-secured). Basically, division of labor between people who are information-rich but cash-poor and people who are cash-rich but information-poor.
3. One interesting point that I think is consistent with my post is that you can replicate many of the risks and consequences of banking by entering into fixed-for-floating interest rate swaps (in which you pay a floating rate and receive a fixed rate - in reality the two are netted, but that is not important for our purposes).
Post a Comment
<< Home