Saturday, January 17, 2004

Brad De Long has a long excerpt from the meeting where Bush appears to have decided to eliminate double taxation of dividends. The clinching argument seems to have been that eliminating double taxation of dividends will cause stock prices to rise (true) and so will improve corporate balance sheets (among other effects) and so will cause increased investment (other things equal) and so is a demand stimulous.

I think there is something general here about how economists can trick non economists with pseudo theory. The phrase in my mind is cherry picking economic theory. Brad has written about this and Krugman has written a lot about this (under heading the accidental theorist).

I think it is possible to help non economists protect themselves.

How do you Know an economist is snowing you ?

A) Never buy an economic argument without numbers attached. A key form of innumeracy is to pay attention to the sign of an effect but not have a clue as to it's magnitude. It is tempting to imagine that if say a policy has 10 good effects and 1 bad effect it is good. This is not arithmetic. I think a defence against an economists snow job is

Economist "eliminating double taxation of dividends will have an imediate positive effect on demand because it will cause an increase in stock prices which will improve corporate balance sheets and cause increased investment"

Pres. "That is a fascinating argument. Do you have a back of the envelope calculation on how many billions more investment there should be in 2003 ?"

Economist "uhm well no"

Pres. "I see. So you don't take that argument seriously at all but just pulled it out of your hat to wow me. You tried to snow me. That's one strike. Two more and you are fired"

B) If anyone mentions only one channel from policy to economic outcomes he is probably hiding something.

Economist "eliminating double taxation of dividends will have an imediate positive effect on demand because it will cause an increase in stock prices which will improve corporate balance sheets and cause increased investment"

Pres. "That is a fascinating argument. Is there any other way in which double taxation of dividends might affect investment. For example is relatively favorable treatment for reinvested profits an incentive to re invest more ?"

a) Economist "There is no other effect at all"

Pres. " You are trying to snow me. That's one strike. Two more and you are fired"

b) Economist "the other effects all imply more investment"

Pres. "Explain"

Economist "Well it is good for going public" (Hubbard could do better)

P You are trying to snow me ...

Economist) well you might imagine that taxing dividends is an incentive to reinvest but, you know, sooner or later the corporation has to pay dividends to get money to its shareholders so taxing dividends is really neutral (this is like argument that corporate income tax has no distortionary effects in the first place but hey wait ...).

Pres. Can't corporations buy back shares ? Why did DEC shares have positive value when their annual reports opened DEC paid no dividend last year, DEC has never paid a dividend, DEC will never pay a dividend ? You are trying to snow me ...


C) Is the economist assuming that economic agents are even more clueless than you ?

P. The last step from the balance sheet to investment makes the implicit assumption that the link between balance sheets and investment is automatic and it doesn't matter why the balance sheet improved. I should know more about this (I am just guessing) but don't balance sheet effects have something to do with banks willingness to lend and investors willingness to buy bonds ? That is something to do with being good for creditors who are after the same pie as shareholders ? Wouldn't something which makes it better for firms to pay dividends make it less good for them to keep liquid assets ? Isn't this a problem for banks and bondholders ? Are you assuming that banks just look at the share price and don't try to figure out why it has changed ?

Well in this case P is clearly flailing.

Economist "Sir you are clearly totally befuddled. I can explain it all very clearly. I will write up a 5 page explanation of why that argument is silly for tomorrow"

Pres. "I see. So you don't take that argument seriously at all but just pulled it out of your hat to wow me. You tried to snow me. That's one strike. Two more and you are fired"

D) is the economist assuming that everyone who is not an economist knows things that he and his smart fellow economists are still debating and have infinite mathematical abilities ?

yes always. Live with it but ask "what would happen if people weren't as smart as you are assuming".


E) So Paul what do you think of this neat argument Glenn just made ?

So recall tool kit for not getting snowed by economists

A) ask for a back of the envelope calculation. No calculation means he hasn't thought about the issue. Doens't want to say means he was counting on your innumeracy. Actually ask for two back of the envelope calculations with explanations of why the results are so different. If the results aren't markedly different he is cheating.

B) Ask for list of important channels of effects of policy. There will be more than one and they won't all point the same way.

C) Read the Lucas critique paper. It really isn't that hard. If the guy is an economist he has considered it and if he doesn't want to say how it is related he is cheating.


E) Finally (obviously) you want debate pro and con so you can use one economist to demystify the other.



Back to history. It seems part of the problem was that O'Neill was not willing and able to debate Hubbard on the balance sheet channel argument. It should have been easy to show that the argument needed more work before it was used as the basis of policy formulation.


No comments:

Post a Comment