Comment by yetanotherphd | Hacker News Reader

Comment by yetanotherphd | Hacker News Reader

Parent Full thread

If I'm right that all stocks benefit equally, then it makes no sense to pick certain companies as "examples". Certainly we wouldn't expect a large differential impact on share prices, and so only a small portion of Amazon's excessive wealth can be explained by the Fed's actions.

Furthermore, it's very hard to separate the direct effect of low interest rates, from improvements in the economy in general. You cite the current price of stocks as an example, but the Fed has had low interest rates since the financial crisis, so you to explain why now in particular is a bubble caused by low interest rates.

On the final point, I am correct about the mechanics of which stocks benefit the most. Suppose instead of high risk stocks, low risk stocks, and bonds, we had apples, oranges and mandarins.

When the Fed buys lots of mandarins, it creates excessive demand for mandarins, and so the price increases. This causes people who were buying mandarins to buy oranges and apples, instead but the total demand for mandarins has still risen. We would expected the biggest price increases to be in mandarins, then oranges, and then apples.

Similarly the biggest price change is in the bond market (which the Fed effectively has complete control over), and then in low risk stocks, and then high risk stocks.

The effect you observe - people being driven from the low-risk to high risk markets by low interest rates - is ignoring the initial price increase that caused this shift.

If I'm right that all stocks benefit equally

You are wrong on this. That's your problem. A dollar today does not equal a dollar tomorrow. Those promising more dollars the day after tomorrow, and those promising a dollar tomorrow will see differential pricing.

I see your point now. However, it's important to note that the Fed's regime of low interest rates is temporary and everyone knows this. So yes, the market will discount investments that pay off in 1-2 years at a low interest rate. But since the low interest rates aren't expected to last, we shouldn't expect huge price increases for stocks whose investments will pay off in 5-10 years.

At the risk of cross-threading the discussion, you need to account for 2nd and 3rd order impacts when trying to figure out where supply and demand will reach equilibrium. In particular, r->0 (asymptotically) is problematic because of somewhat opaque issues with DCF pricing. In particular, R(r) where R is the market discount rate and (r) is the risk free rate. Lets assume the fed is manipulating (r) and that R(r) is directionaly (+). That is fair enough. But R is also impacted by exogenous preferences setting (for risk, weath effects, substitution effects). While these preferences may be stable in theory at "normal r" (~econ 101 assumption), they are empirically not independent of (r) as r-->0 and so are not stable at all in our test case. In fact, as (r)-->0, it becomes more and more rational to take on extreme risk (especially so if (r) is temporarily/ artifically low. After all, you are playing with "free money". Now, lets assume that you have shifted gears in this way. Lets further assume you have a choice of two investment classes: (a) Low risk and (b) High Risk. Its clear you will weight your portfolio towards (b), purely through the change in preferences. This change in <preferences> will change R(r) and show up in your DCFs (ie, when implied by market pricing). In other words, the only way 'DCF' will make sense is when you plug in R'(r)=|=R(r). But now lets consider the subset of (b) in isolation. Given the result above, R'->0, stocks that have non-transparent, distant income will become relatively more attractive. That is to say, the current prices will look too low. So, you will overweight these stocks (bidding up prices). But not only will you in general bid up all assets (due to little r) you will bid up disporportionately equity vs debt and furthermore non-transparent risk equity vs lower-risk equity. As a result, you <should> expect huge price increases for stocks whose investments will pay off in 5-10 years.

Or at least, so the theory goes.

I think you are now the one doing the over-thinking.

You are applying a model with fixed r, which is wrong because the current low interest rates are temporary.

Forget the interaction between interest rates and risk preference for the moment. Isn't it the case that if interest rates are certain to return to normal in 2 years, then the values of an investment that returns in 5 years, and an investment that returns in 2 years, will be equally affected by current low interest rates?

You are applying a model with fixed r

No, this is not true. The whole point is that R=R(r). Its just that the relationship is ~chaotic, not linear. And it gets messy as either r or R --> 0. The fact that preferences are recursive on (r) should not be a surpirse, but I don't think mainstream Economists generally go there.

On the topic of "temporary" vs Uncertain changes in (r), they are not the same. One is risk, the other is uncertainty. In general risk is priced a certain way (reasonably efficient), uncertainty is something you have preferences for (ie, you dont "price" it in the normal sense) and can only model indirectly in the first instance.[1]

[1] eg, http://economistsview.typepad.com/economistsview/2010/06/ris...

> However, it's important to note that the Fed's regime of low interest rates is temporary and everyone knows this.

The general consensus just a few short months ago was that the Fed would start tapering in September. It did not happen. There are credible observers and market participants who not only believe that tapering is not coming any time soon but that the Fed's asset purchases will actually increase.

I don't necessarily agree, but I do believe that the Fed has positioned itself between a rock and a hard place, so I personally wouldn't make too many assumptions about timing and magnitude of increases or decreases in asset purchases.

> So yes, the market will discount investments that pay off in 1-2 years at a low interest rate. But since the low interest rates aren't expected to last, we shouldn't expect huge price increases for stocks whose investments will pay off in 5-10 years.

You make the mistake of assuming that investors are entirely rational. The empirical evidence clearly demonstrates that this is not the case. You should also consider that many (and perhaps the majority) of the participants in a good number of markets aren't really "investors" as most people define the term.

John Williams, the President of the SF Fed, recently addressed asset price bubbles in a presentation[1]. It is well worth a read, and I would point you to the following:

"For monetary policy, one implication of theories with endogenous asset price bubbles is that the time horizon over which policy affects the economy may be longer than typically thought. In particular, the policy response to cyclical movements in economic activity and inflation may have effects on investor beliefs and the behavior of asset prices that reach well into the future."

[1] http://www.frbsf.org/our-district/press/presidents-speeches/...

Reply on news.ycombinator.com