Paul Krugman has an interesting post, where he pushes back against the conventional view that micro is more successful than macro. Here’s Krugman:
So let me talk about three things:
The unsung success of macroeconomics
The excessive prestige of microeconomics
The limits of empiricism, vital though it is
I think Krugman’s basically right, although I’d emphasize some different points.
Krugman spends a lot of time on the IS-LM model, which he finds useful and I do not. He emphasizes the fact that IS-LM allowed him to avoid falling into the trap of expecting QE to lead to high inflation. Rapid increases in the money supply are often inflationary, but not when interest rates are near zero.
I prefer to think in terms of money supply and money demand when trying to explain movements in nominal aggregates. For instance, macroeconomists who understand monetary theory can easily explain why inflation has averaged 2% since 1990, whereas non-economists and economists with non-monetary models have no explanation at all. And no, the following does not constitute an “explanation”:
1. Congress
2. Fiscal policy
3. A miracle happens!
4. 2% inflation for 28 years
Macroeconomists also have a pretty good model of the business cycle. The combination of nominal shocks and sticky wages/prices results in short term movements in employment and output that are highly correlated with short run movements in NGDP. In the long run, wages and prices adjust and the economy goes back to the natural rate. These theories are not perfect (hysteresis is not well understood), but they are pretty good.
So why can’t we predict business cycles? For the same reason that microeconomists cannot predict individual prices—economic theory (such as rational expectations and the EMH) suggests that many types of predictions are almost impossible.
Macroeconomics is also pretty good at explaining long run growth. It seems to be based on 5 factors (human capital, physical capital, technology, natural resources, and good institutions.) Yes, the final one is a bit vague, but probably involves relatively free markets, property rights, non-excessive taxes, etc.
Microeconomists have one highly successful model—perfect competition. But most markets are not perfectly competitive. As soon as you move to imperfect competition, micro is just as ambiguous as macro. Minimum wages? Who knows if they reduce employment? Health care economics? A huge mess, where economists cannot even reach a consensus on the key issues that need to be focused on.
Microeconomics might be a bit more successful in some ways, as the problems being considered are often less complex. As an analogy, macrophysics (earthquake prediction, weather forecasting, etc.) is less precise than microphysics (a rocket’s trajectory through outer space.) But less complexity is the only sense in which micro is more “scientific”.
The law of large numbers allows certain parts of macro (inflation, business cycles, etc.) to actually be better understood than some parts of micro (health care, antitrust, etc.) Inflation is simply about money supply and money demand. In contrast, health care is about barriers to entry, information asymmetries, moral hazard, adverse selection, intellectual property rights, cross-subsidization, mandates, and many other market imperfections.
A macro perspective often gives a more complete picture of the economy. Thus you might (wrongly) assume from micro that immigration would reduce wages. More immigrants means more supply of labor, which means lower wages. Of course this isn’t actually a prediction of micro, but it’s an easy trap to fall into (read commenter Ben Cole). A macroeconomist would notice that America has similar wage levels to Canada, despite roughly 10 times as much immigration. Then she might notice that immigration also increases the demand for labor, indeed at roughly the same rate that the supply of labor rises. This idea is sometimes called the “macro foundations of microeconomics.”
Krugman’s also right that data are not enough; we need to combine data with theory to make sense of the world.
PS. A quick follow-up on my previous post. It was southern Orange County (Trump country) that opposed the nice new airport; the more Democratic northern part of the county supported the project. In other areas of the country, it could easily be the reverse. Remember that next time someone asks you, “why can’t we have nice things?”
READER COMMENTS
E. Harding
Sep 22 2018 at 4:12pm
That actually helps the immigration restrictionist case; “look, immigrants lower our wages AND raise our prices as well”. The real reason immigration does not lower wages in general is the same reason automation doesn’t, that is, due to the complementarity of labor, which is much more a micro concept than a macro one.
Scott Sumner
Sep 22 2018 at 6:17pm
Harding, Sorry, I don’t follow that at all. My claim is that it doesn’t lower (real) wages, because both the supply and the demand rise by roughly equal amounts.
Lorenzo from Oz
Sep 22 2018 at 8:21pm
Canada is also much more selective about its migrants than the US, given it has very little illegal migration. Australia is in the same situation. Labour is heterogeneous, so inferring from size is perhaps a little fraught.
Hazel Meade
Sep 24 2018 at 11:39am
People keep saying this about Canada, but if you want to immigrate to Canada permanently you can get a visa approved in under a year, whereas it can take 3-5 years in the US, and longer (10 years or more) in family sponsorship categories.
I think the main difference is that the point system Canada applies means that many more family sponsorships are disapproved. In the US there is a quota and a very long waiting list – nobody gets disapproved, they just wait in line forever. Meanwhile it is much easier to get an skilled worker visa in Canada because you don’t have to go through the labor certification process that the US has. So if you’re coming to work, it’s easier to immigrate to Canada.
P Burgos
Sep 24 2018 at 12:43pm
I think that that another big difference between the US and Canada/Australia is that neither Canada nor Australia have a long land border with a much poorer nation that has very strong organized crime syndicates and a long history of border smugglers. So it is relatively easy for Canada/Australia to identify and apprehend people who violate the nation’s immigration laws, because there aren’t too many people who do.
Floccina
Sep 25 2018 at 4:02pm
Yup selection effects, when Canadians pat themselves on the back about how much better their immigrants are doing, ask them how well their First Nations and Inuit are doing (Same with Denmark and their Inuit.) Our Hispanics are doing great considering where they come from.
Lorenzo from Oz
Sep 22 2018 at 8:29pm
I am a bit less impressed by the understanding of long term economic growth than you seem to be–see arguments over what caused the Great Enrichment and why it started in Britain.
Also, institutions don’t seem to matter much for economic growth until we get to the Great Enrichment. And the literature has been rather too focused on state institutions (though the work of Avner Greif and others is improving that). Thus the revitalisation of clans in China seem to explain much of why China has taken off even though its state institutions do not seem to conform to the “property rights and rule of law” model.
Philo
Sep 22 2018 at 9:48pm
And good institutions, including “free markets, property rights, non-excessive taxes,” themselves depend on the public’s having a favorable attitude towards commercial activity *per se*.
E. Harding
Sep 22 2018 at 9:55pm
Wouldn’t real wages rise more if they didn’t contribute demand for the products of labor (and, thus, didn’t raise prices)? Don’t machines contribute more to higher real wages when they produce more than they consume?
Rajat
Sep 22 2018 at 10:06pm
Just seeking some clarification on the following:
I am left uncertain as to whether you regard IS-LM as a non-monetary model (I assumed it was – after all, the LM curve defines money market equilibria). If IS-LM is a monetary model, then while you prefer a different model, that does not seem crucial to distinguishing what kind of economists have an explanation for 2% inflation since 1990. Or are you saying that Krugman has no explanation for 2% inflation due to this reliance on IS-LM? Assuming the former, I wonder if readers would find it helpful for you to briefly refer to the non-monetary-type models you are criticising – eg do you mean modern New Keynesian models generally or just DSGE?
Scott Sumner
Sep 22 2018 at 10:41pm
Lorenzo, You said:
“Thus the revitalisation of clans in China seem to explain much of why China has taken off even though its state institutions do not seem to conform to the “property rights and rule of law” model.”
I have to strongly disagree with that. China’s economic growth is almost exactly what you’d expect of an East Asian country with those property rights and markets. China’s vastly more market-oriented than in 1980, but still has a long way to go, which is why it’s vastly richer than in 1980, but still has a long way to go. The standard growth model explains China extremely well.
Harding, I think you are confusing demand for labor with demand for goods.
Rajat, I’m talking about MMT models, Fiscal Theory of the Price Level, etc. The IS-LM model does include the money market, but I prefer a simpler money supply/money demand model.
Benjamin Cole
Sep 23 2018 at 12:50am
OT, but Beckworth’s latest post about the Fed actually forecasting a negative yield curve next year is worth some deep thinking, or at least abject fear and depression.
As an aside, I wonder if China and Singapore prove anything about anything. Free markets?—or some of the most regulated economies on the planet?
The joke is that Far East governments are so pro-business, they often decide to go into business themselves, and to ensure profits for their business constituents.
In Hong Kong, property developers and lenders work closely with government to ensure each new development is needed and works. Oh, gee…
You only learn this if you examine Hong kong development patterns, and get in there pretty deeply with people who know development in Hong Kong…it is not a “macroeconomic” topic.
I think people who go to public policy schools might understand Far East economies better than macroeconomists, although there are blind sorts all around.
Singapore much the same.
This Far East blending or public and private is not wrong or right. Far East economies and cultures have a different take: government and business is a marriage with the offspring called commerce.
Mike Hammock
Sep 23 2018 at 8:38am
I would add auction theory and mechanism design to micro’s successes.
Scott Sumner
Sep 23 2018 at 11:05am
I agree.
Scott Sumner
Sep 23 2018 at 11:05am
Ben, You are not well informed about East Asia; Hong Kong and Singapore are the two most free market economies in the world according to every ranking I’ve seen. You can’t beat systematic studies with a few questionable anecdotes.
robc
Sep 25 2018 at 8:41am
Rapid increases in the money supply are often inflationary
Especially if you define inflation == monetary inflation. Then you can even drop the “often”.
Mark Z
Sep 25 2018 at 10:50am
I think micro is inherently more reliable because mircroeconomists (excepting those who study extremely rare or obscure phenomena) have far more data than macroeconomists do, so we can far better assess what clearly works and what doesn’t with far less controversy than exists in what works in macro modeling. I think Arnold Kling writes about this a lot. There just aren’t enough independent data points in macro analysis – years, cycles, etc. – to approach the level of confidence we can have on most micro topics.
Your contention on the inherent difficulty or impossibility of making short run macro predictions actually might make a good analogy to climate science: it’s generally pretty good at predicting long run trends; not very good at short run trends (in the case of climate, because of lots of short run variability), but honest climate scientists will not profess to be able to predict short run trends; nonetheless some people declare the field a failure for failing to predict short run trends because that’s what they care most about; and some scientists get cocky and overshoot and profess, on behalf of the science, to be able explain or predict short run trends on behalf of the science better than it really can.
Scott Sumner
Sep 25 2018 at 11:16am
robc, Yes, but that’s not a useful definition of inflation. It’s not helpful to say that inflation is caused by inflation.
robc
Sep 25 2018 at 11:30am
I may be wrong, but wasn’t that the standard definition of inflation used by economists until about 50 years ago? Didn’t “inflation” switch from “monetary inflation” to “price inflation” in general economic usage about the 60s?
Maybe my timing is wrong on that, not sure.
It is a useful definition of inflation if you don’t consider prices to be a measure of inflation.
You said “Inflation is simply about money supply and money demand.” If you change inflation from price inflation to monetary inflation, you can simplify even more and eliminate the last 3 words.
robc
Sep 25 2018 at 11:37am
Also, I do think it is still useful if you are talking about something like M2. Then there are two major factors in the inflation, change in M0 and change in velocity of money (I know I am simplifying, but those would be the two big factors). So, it would be possible for something like QE to not be inflationary if it matched a fall in the velocity of money.
I am not an economist, but this seems like a better measure of the inflation rate to me than the cpi:
https://fred.stlouisfed.org/graph/fredgraph.png?g=llGn
Scott Sumner
Sep 25 2018 at 10:54pm
robc, No, the quantity of money was not the standard definition of inflation (although it was occasionally defined that way.) Indeed early economists believed that inflation was actually caused by increases in the money stock. Read Irving Fisher.
As for M2, I don’t see why that matters. Suppose you told someone that inflation was 10% back in 2001, because M2 grew at 10%. What would that fact mean to the average person? How would that help them with their life? Wasn’t the 2% rise in the CPI more meaningful? (Even though the CPI is also flawed.)
If we define inflation as money growth, then what do we call the rate of increase in prices?
robc
Sep 26 2018 at 10:26am
Asking the last question first: purchasing power?
Although the increase in prices would be the inverse of the change in purchasing power. But it measures the same thing.
To me, that would be a better term, as it better describes all the things that change prices, such as productivity gains.
In some sense its just semantics, it doesn’t matter what we call it.
As far as the previous paragraph, it tells the average person how much of the money supply their paycheck represents. In a fiat money system, that seems important to me, as the total amount is arbitrary.
If we pull a reverse Zimbabwe and arbitrarily add 3 zeros on the end of our dollar, I want my paycheck getting 3 zeros too.
If our money is backed, I wouldn’t care as much; if someone triples the money supply with a big gold find, good for them. It might be a disaster for my purchasing power, but it was a free market in gold hunting that I missed out on. I would still have the same amount of gold or silver or palladium or whatever I had before.
LK Beland
Sep 25 2018 at 3:15pm
“A macroeconomist would notice that America has similar wage levels to Canada, despite roughly 10 times as much immigration.”
I don’t quite get this statement. Canada’s net migration rate is twice that of the US.
Scott Sumner
Sep 25 2018 at 10:47pm
LK, I’m saying that in total the US has received roughly 10 times as many immigrants. If all that mattered was the supply of labor, then wages should be far lower here than in Canada.
Mark Z
Sep 27 2018 at 1:22am
So, you’re using the absolute number of immigrants rather than as a fraction of population?
Is there some reason I’m missing why 10 million immigrants moving to a country of 10 million would have the same effect on that country’s wages as 10 million immigrants moving to a country of 100 million?
LK Beland
Sep 27 2018 at 1:15pm
“I’m saying that in total the US has received roughly 10 times as many immigrants.”
It’s actually 5, not 10.
Mark Z says: “Is there some reason I’m missing why 10 million immigrants moving to a country of 10 million would have the same effect on that country’s wages as 10 million immigrants moving to a country of 100 million?”
Exactly. I think that all commentators agree that the arrival of 100,000 low-skill immigrants in California will have a different effect on wages than the arrival of 100,000 low-skill immigrants in Wyoming.
A better statement would be: Switzerland has triple the immigration rate of the US, but has higher wages.
Another good statement: Canada’s immigration rate is twice that of the US, and its low-skill wages and prime-age employment rates are higher than the US’.
Comments are closed.