The real “long and variable lags” problem

One of the most overused cliches in monetary economics is that policy affects the economy with “long and variable lags.”  In fact, policymakers should be targeting price level or NGDP expectations, not the actual price level or NGDP.  And monetary policy affects expectations with a zero lag.  Even worse, future expected changes in NGDP have a powerful effect on current NGDP, so the so-called “policy lag problem” is a phony issue even if we did want to control actual NGDP.

There is one area where long and variable lags do exist, however, the supply-side effects of fiscal policy.  Consider an example where doctors require so much education that students won’t be willing to go through the arduous process of medical school and residency unless they can eventually earn twice as much as other white collar workers.  Let’s suppose that in a zero-tax world white collar workers earn $100,000 before and after taxes, and doctors earn $200,000 before and after taxes.

What would happen if the government suddenly put a 50% tax rate on incomes above $100,000/year?  Initially the effect would be small.  Doctors might work somewhat fewer hours, but since their medical education is a sunk cost this tax increase probably wouldn’t cause many to go back to school in order to learn accounting.  Over time fewer students would go to medical school, and in the long run a new equilibrium would be reached where doctors earned $300,000/year, i.e. $200,000 after taxes.

What can we learn from the previous example?

1.  It takes a very long time for the full effects of supply-side tax changes to fully impact the stock of physical and human capital.

2.  A tax that may appear to be paid by “the rich” might in fact be paid by the customers of the rich.

I am not saying that there aren’t some immediate effects of tax changes.  Nor am I denying that some taxes affect the rich more than others.  Part of wealth comes from IQ and ambition and luck, not investments in human capital.  Rather I am saying that the sort of government data often used to discuss these issues is nearly useless, as it measures the legal incidence of taxes, not the economic incidence.

Here is another example.  Let’s suppose that one state puts a high tax on individuals making more than $100,000/year.  What is the effect?  There are many possibilities.  One (which we just considered) is that skill levels in the state gradually decline.  Another possibility is that people leave the state, and move to lower tax states. But once again, this will occur very gradually over time.  A doctor with an established practice many not want to pick up and leave New Jersey, but if his son or daughter is just coming out of medical school, the tax increase may push him or her to a lower tax state.  And of course physical and cultural amenities can also influence location decisions, which is why New York City and California can get away with higher tax rates than Michigan, Ohio, upstate NY and Texas.  The strongest tax effects are people moving from boring high tax states in the rust belt to boring low tax states like Texas.

Because the lags in tax policy can stretch over decades, I am very suspicious of time series studies of taxes.  The best evidence comes from comparing incomes (PPP) in different developed countries.  This picks up differences in hours worked, and also differences in productivity associated with variations in human and physical capital.   BTW, the fact that productivity per hour in the US, France and Germany is fairly similar, actually reflects poorly on the two European countries.  You’d expect productivity to be higher where hours worked are lowest, as the least productive workers are those most likely to be unemployed.  This may explain some of the recent divergence in productivity between Germany and the US.   As employment in the US has fallen closer to German levels, and output in Germany has fallen faster than output in the US, productivity in the US must have risen rapidly relative to productivity in Germany.

I plan to do a series of posts on taxes.  The goal is to show the advantages of a progressive consumption tax (plus Pigou taxes and perhaps land taxes), and also the fact that bar charts showing the share of taxes paid by each income quintile are basically worthless.



12 Responses to “The real “long and variable lags” problem”

  1. Gravatar of Doc Merlin Doc Merlin
    20. July 2010 at 12:50

    Yay! Microeconomic arguments! Yay!

    er… um… good post!

  2. Gravatar of Joe Joe
    20. July 2010 at 14:09

    How do you explain the fact that in the post war era we had massive marginal tax rates yet greater economic growth and prosperity? Thats the typical argument from the left….

    Can it be empirically/research-wise proven that high marginal tax rates are bad for productive behavior, or is this argument based simply on theory?


  3. Gravatar of Thorfinn Thorfinn
    20. July 2010 at 14:12

    What are you thoughts on a land tax? One of your hotspots, Hong Kong, has gov ownership of land and rents it out to pay for a good chunk of revenue, and Martin Wolf has argued for expanding this model, Henry George-style.

  4. Gravatar of Lord Lord
    20. July 2010 at 15:35

    If expectations are instantaneous then why could Volker not end inflation overnight as he tried to do so well? Expectations require believability that can only be established over time.

    “You’d expect productivity to be higher where hours worked are lowest, as the least productive workers are those most likely to be unemployed.”

    Incorrect. People are paid according to their productivity so there is no more likelihood for less productive workers to be unemployed. Rather, one expects higher productivity from fewer hours because productivity declines with longer hours.

  5. Gravatar of scott sumner scott sumner
    20. July 2010 at 16:00

    Doc Merlin, Thanks.

    Joe, I did a bunch of posts on that a few months back.

    1. The post-war period saw rapid growth almost everywhere, from Mexico to Brazil to the US to the Soviet Union to Romania to Italy to Greece. It doesn’t tell us much about whether the economic model was actually very efficient.

    2. The US economy actually improved after the Dems cut the top rate from 90% to 70% in 1964.

    3. Most people did not face high MTRs during the post war years, only a tiny few.

    4. The very rich had all sorts of loopholes, such as 60% exclusion of capital gains.

    So it wasn’t a good system, but it did limited damage, and the economy actually did better after rates were cut far below 90%. And those were fast growth years for economies (Russia) that no one now would want to emulate. It was also rather pointless, as it collected little revenue.

    Unfortunately things are complicated, and simple sound bites win arguments.

    Thorfinn, I think land taxes have a lot going for them. From what little I know about Hong Kong, the government is too restrictive on land use. This keeps housing prices high, and lowers living standards. But that’s a separate issue from the land tax itself.

    Lord, The market was wise to not initially trust Volcker. He started off in 1979 with a tight money policy, but in mid-1980 switched back to easy money and inflation started rising again. Only after Reagan took office and gave him political cover did he actually get serious about inflation (in mid-1981).

    I disagree on productivity for two reasons. In Europe the long term unemployed are those whose productivity is below the minimum wage, and/or those for whom living on public assistance is better than the alternatives. Both groups tend to be relative low-skilled. The low-skilled immigrants who tend to be employed in the US are often unemployed in Europe.

  6. Gravatar of Morgan Warstler Morgan Warstler
    20. July 2010 at 17:19

    Simply because you need to read it:

    The macroeconomic consequences of this shift toward low-equity homeownership are visible in research from the Federal Reserve that examines the assets and liabilities of U.S. households. In the first quarter of 2001, U.S. households’ home equity stood at $7.7 trillion, or 61 percent of the value of all residential real estate. By the third quarter of 2008, it had declined to $7.6 trillion, even as outstanding mortgage debt increased by $5.6 trillion over the same period. By the first quarter of 2009, home equity was $1.35 trillion lower than it had been in 2001. Put another way: Despite the housing boom, the portion of residential real estate actually owned by households declined. This means that the increase in homeownership rates (and the subsequent rise in housing prices) was entirely debt-financed.

  7. Gravatar of scott sumner scott sumner
    21. July 2010 at 04:33

    Morgan, Sorry, I am not interested in the housing market. I am a macroeconomist.

  8. Gravatar of Lucas Lucas
    22. July 2010 at 00:33

    Wow, a self-described libertarian supporting Pigouvian taxes and georgism [1].
    Actually, it’s a pleasure to read smart arguments from intellectual adversaries, given the dearth of enlightened right-wingers.
    Regarding progressive consumption taxes, what’s your opinion of the “Efficient Taxation of Income”? [2]
    Do you think that taxes/fees on natural monopolies should be higher? What about the Tobin tax? [3]
    What would a libertarian do with the raised revenue?


  9. Gravatar of Rick Rick
    22. July 2010 at 06:02

    I would love to hear your thoughts on the land tax. It has always seemed like one of the most sensible policies not generally championed by many economists.

  10. Gravatar of scott sumner scott sumner
    23. July 2010 at 05:44

    Lucas, Thanks, but there are plenty of right-wingers much more “enlightened” than I am. Check out blogs like Marginal Revolution, or Becker-Posner, or Econlog.

    I don’t know enough about that plan to have a firm opinion. It looks like a consumption tax plus a tax on existing capital income. Over time it would morph into a pure consumption tax. At least I think that is what it is, but I can’t be sure. The description is too sketchy for me to follow exactly how it works.

    I oppose the Tobin tax. It seems pointless.

    I don’t favor increasing total tax revenue, so I have no views on how the extra revenue should be spent. I think government should spend about 15% – 20% of GDP, which is much less than the 35% we actually spend. So if we implemented these taxes, I would eliminate other taxes like the income tax.

    Rick, I think it could be sensible, if done right. Much of the property tax in high cost places is actually a land tax. So in a way we already have it, but of course in a less efficient form that also hits building improvements.

    My guess is that many economists don’t like it because they omit the third step of the following thought process:

    1. It looks perfect at first glance
    2. At second glance there are some problems
    3. At third glance there are much worse problems for other taxes.

    Many economists that I talk to seem to stop at step two.

  11. Gravatar of Doc Merlin Doc Merlin
    23. July 2010 at 07:45

    The housing market is about half of american’s savings and intimately involved in just about every single downturn, you should look into it more, for its macroeconomic significance.

    Quite a few libertarians are such. I am not one of them, but Greg Mankiw is one, and so is Tyler Cowen

  12. Gravatar of ssumner ssumner
    24. July 2010 at 16:24

    Doc Merlin, I’m not totally convinced on either point. I think business investment usually is more cyclical, although this time housing was important. And where does the 50% come from? That seems a bit high, espcially if you include all forms of saving (pensions, buiness saving, etc.)

Leave a Reply