Let’s try one more time, with the dollar/yen forward exchange rate. I’d like to make the following assumptions. It doesn’t matter whether you think these assumptions describe the real world; I’d simply like you to consider them as a hypothetical. When we’re all done, we’ll think about what it means.
1. Let’s assume the BOJ is determined to adopt a very tight money policy, over the next 30 years. This policy will be so tight that the yen will end up valued at 50 to the dollar, more than double its current value.
2. This very tight money policy causes very low inflation and very low NGDP growth in Japan.
So far interest rates don’t enter the picture, indeed interest rates need not even exist—imagine a world with no debt. I’m trying to make the appreciation of the yen into the fundamental shock, from which everything else flows.
3. Now let’s add interest rates. Because of the ultra-low expected inflation, and the ultra-low expected NGDP growth, nominal interest rates in Japan are more than 200 basis points below nominal interest rates in the US. These low rates are caused by a tight money policy that leads to yen appreciation.
I’m still assuming the tight yen policy that leads to yen appreciation is fundamental, and everything else is an effect of that policy.
4. Now let’s assume that the US and Japanese debt markets are very deep and liquid, and the 30-year forward yen contract is very lightly traded and not very liquid at all. Let’s also assume that the forward premium on the yen is linked to the interest rate differential according to the covered interest parity theorem, although the theorem doesn’t work perfectly due to various market imperfections caused by regulations. It’s roughly true.
I’m still assuming the tight yen policy that leads to yen appreciation is fundamental, and everything else is an effect of that policy.
Now let’s take stock of where were are. Thus far, I have NOT claimed to describe the real world. I’ve described a scenario where, by assumption, the huge forward premium on the yen drives the interest rate differential. Quite possibly, this imaginary scenario has nothing to do with the real world.
But here’s the problem. Not one commenter has given me a single fact that would lead me to conclude that this imaginary scenario does not in fact describe the real world. Note, for instance, that I assumed that the two bond markets are highly liquid and traders focus on the interest rate spread. I assumed the forward yen is lightly traded, and hence considered peripheral in the world of finance. But I’ve also constructed an example where, by assumption, that difference in liquidity between the two markets has no bearing on causality.
So what would count as evidence against my imaginary scenario? Perhaps you could convince me that while the 30-year forward yen is 50, traders actually expect the yen to be trading at 105 in the year 2048. And investors continue to buy low yield JGBs in any case, because of market segmentation, or some other reason. So the differences in interest rates are unrelated to differences in inflation, etc. If you offered that sort of explanation, and backed it up with evidence, I would be persuaded. But I’m not seeing people do that. Until then, I’m going to assume the causality goes from an appreciating yen to a situation where Japanese interest rates are lower than American interest rates.
PS. The “carry trade” may partly explain why people disagree with me, but carry trades suffer from the “peso problem”, so I’m not convinced the carry trade will “work” going forward. If Japanese inflation stays well below US inflation (as I expect), then the carry trade will break down at some point.
PPS. Financial variables may or may not be linked to macro events. The 1929 stock market crash seems to have been linked to fears of depression, while the 1987 stock market crash seems to have been sort of random. You can view my claim here as being that the 1929 case is more typical. Asset prices move based on shifting expectations regarding economic fundamentals. Even if a forward exchange rate market did not exist, I’d claim that expectations of the future spot rate drive the interest rate differential.