Arnold Kling  

Models Vs. Hand-Waving

Timothy Taylor on NCH... Don't Tell Tyler Cowen...

Nick Rowe writes,

the answer to the question "what is the effect of a 1% increase in perceived risk on government bonds?" is exactly the same as the answer to the question "what is the effect of a 1% increase in expected inflation?". At least, in this model, given my assumptions. Both result in the same 1% fall in the real risk-adjusted rate of interest on private saving and investment and the same rise in AD and real income.

He wonders why Keyensians are not cheering a political deadlock which would raise the risk premium on government debt.

I think most Keynesians would wave their hands and say that "A breakdown in trust in the financial system would be bad for aggregate demand." I think that would be a better answer than saying that, "Well, in my model, an increase in the risk premium on government debt is expansionary, so that must be the right answer."

A less satisfying answer is to say that "All securities are priced off of government debt. If the interest rate on government debt goes up, then all other rates go up." I do not think that is what a model would say. If it is right, it is hand-waving.

Which suggests that hand-waving captures a truth that is not in models.

Comments and Sharing

COMMENTS (7 to date)
Philip George writes:

"If the interest rate on government debt goes up, then all other rates go up."

The assumption seems to be that this is a bad thing and will hamper growth.

But this is the classic error of viewing a situation from the standpoint of demand alone. Looked at from the supplier's standpoint, high interest rates are not a bad thing at all.

Both the logic, and the evidence from 2001 to 2011, shows that it is quite the opposite: it is when interest rates fall to very low levels that bank lending dries up.

See for more arguments and numbers.

Chris Koresko writes:

Arnold Kling: "...hand-waving captures a truth that is not in models."

If you want to predict economic performance, you use a numerical model. That model must leave out non-quantifiable factors which may substantially affect the real outcome. In the case of bad policy, the model is systematically biased in the direction of making optimistic predictions.

If policy makers rely on numerical models, those models may encourage them to continue bad policies by predicting good outcomes.

robbl writes:


you seem to be quoting this approvingly, but I can't believe that you believe it. You are too sensible. I know cause I read this blog on occasion.

A 1% increase in the risk premium for US government bonds would a gigantic tsunami in the financial markets that would roil things around for years.

A 1% increase in inflation happens all the time with hardly a murmur.

How much hand waving do you have to do to equate the two?

Dale writes:

The answer depends on what you think a 1% change in interest rates represents. Does it represent a decrease in demand for safe assets or does it represent government assets getting less safe?

If it is the first, then it is not bad, increased interest rates have the same effect as inflation. I.E. it allows the market for safe assets to come closer to equilibrium which similarly forces the rest of the economy closer to equilibrium.

But if it is the second then it makes the problem worse, because instead of allowing the market for safe assets to get closer to equilibrium it widens the gap. I.E. people still wants tonnes of safe assets and there are fewer safe assets around. This widens the gap between demand and supply which creates an even larger shortfall in GDP.

Now, the second explanation doesn't make much sense because in this model, assets can only be "safe" or "unsafe", demand isn't shared among the two classes(some amount of substitutability yes), and so either a 1% increase in interest makes things better or it causes a horrible calamity as all treasuries become unsafe and we get an even larger excess of demand for safe assets.

But the second doesn't make sense either, why would dollars leaving demand treasuries necessarily go into goods/services? So let us take the first explanation, a reduction in demand for safe assets, but modify the model a bit to allow for that demand to move into different places in the economy.

Problem: A reduction in demand for treasuries implies an increase in demand for other aspects of the economy. Would we expect that people, no longer able to get the assets they want because they don't consider them safe enough would start buying unsafe assets or goods/services?

Probably not. If they did, then it would be fine, the reduction in demand for treasuries would mean an increase in demand for the rest of the economy, which would increase GDP. But this is not a reasonable assumption, because they are not buying treasuries because treasuries became more risky.

Instead it is more likely to say that they will increase their demand for cash, or other stores of value that are considered safer than treasuries. And if this happens and those assets aren't also oversupplied/underdemanded then you have made the problem no better. Every dollar you drop in excess demand for treasuries gets increased in excess demand for other safe assets like cash and gold.

The model may not allow this specific situation, but the models they're using probably don't need to under normal circumstances.

Indy writes:

Is it actually true that there are no AAA Corporate Bonds in the world that get better rates than their respective sovereigns?

Dale writes:

Having thought more about it, the first explanation(modified) breaks the assumption that the fed can perfectly supply the cash at whatever nominal interest rate it wants. Having read the comments at WCI i am not sure that my explanation makes sense.

Floccina writes:

I find it interesting to think about what people would do if they had less confidence in Gov. bonds (including paper money which seems to me to be a bearer bond that pays no interest). I think that they would hold less of it and perhaps buy land or equities or money saving assets (like home insulation, cars that get better mileage etc.).

Comments for this entry have been closed
Return to top