« Assorted links | Main | How to choose an apartment »
Trade and inequality, revisited -- Rooftops edition
Another way of investigating the relationship between inequality and trade with poor countries implies that China may actually help the poor, suggests new work from University of Chicago economists Christian Broda and John Romalis.
Instead of focusing purely on what's produced outside of the country, Broda and Romalis turn their attention to an interesting but obvious relationship between imports and consumption within our border: The goods exported by poorer countries are typically consumed by lower-income Americans. Our typical methods of quantifying inequality, however, don't take this into account.
At the same time, inflation in the price of these goods has fallen behind inflation in services, which make up a greater portion of what wealthier people buy. Taken together, these trends imply that official measures may be overstating the rise in inequality.
Looking at trade data between 1994 and 2005, Broda and Romalis construct inflation rates for different income groups and find that rates for the richest outpaced rates for the poorest by about 4 percent over the period. Since income inequality between the top and bottom 10 percent of earners grew by about 6 percent, the different inflation rates among income groups wipes out about two-thirds of the rise in inequality.
The emphasis in that last sentence is mine. It continues:
China's role in this new way of analyzing inequality is large, accounting for about 50 percent of the total reduction.
And scream this part from the rooftops too [how do you scream a parenthesis?]:
(A very interesting aside. Broda and Romalis also find that the poor are more likely than the rich to buy newer goods. Because of the lag in how quickly the CPI tracks new products, the researchers argue that once this "new goods bias" which serves to keep official inflation rates higher than they actually are since newer goods are typically cheaper, is factored out, inequality between the rich and the poor between 1994 and 2005 may not have changed at all.)
Here is the link. Again, here is the Broda and Romalis paper. If this holds up it is big, big news and we must revise many claims that have been made about inequality, trade, and China.
Posted by Tyler Cowen on April 25, 2008 at 10:54 PM in Economics | Permalink
Comments
Isn't this very related to the idea that consumption inequality is a lot lower than income inequality?
Posted by: jsalvati at Apr 25, 2008 11:21:50 PM
Figure 10 indicates that the inflation bias is nearly constant for incomes between $15,000 and $100,000 a year where the vast majority of full time workers are. It is not the poor with incomes less than $15,000 that I picture having lost well paid manufacturing jobs to China, but the median wage earners nearer $30,000.
Posted by: joan at Apr 25, 2008 11:42:52 PM
Is the Chinese government's dollar peg (an American middle-class/poor subsidy) taken for granted in the study? Would a real free-trade regime with a floating dollar not backed by petrodollar hegemony or mercantilism produce similar results?
Posted by: ideogenetic at Apr 26, 2008 12:35:27 AM
I've seen Broda present this paper. I direct all readers to the headline results in Table 7B. The "two-thirds" and "almost wiped out inequality" conclusions come from the 1994-2005 line (i.e. 3.7/5.7 and 5.5./5.7). But the authors are not highlighting the results for 1999-2005, during which the corresponding adjustments are just 24% (1.7/7.2) and 46% (3.3/7.2). I really like the paper, but the strength of the conclusion certainly depends on the time period.
Posted by: DRDR at Apr 26, 2008 12:44:55 AM
The problem with being poor in 21st Century America is not that you can't buy enough stuff, it's that you have to hang around with other poor people.
Posted by: Steve Sailer at Apr 26, 2008 12:53:33 AM
Tyler,
What exactly about this study didn't you already know intuitively? Didn't you already know that this country has been buying boatloads of cheap gizmos from China on credit?
Posted by: Steve Sailer at Apr 26, 2008 12:57:24 AM
Differing inflation rates for differing income groups, what? How does that work? Are they calling any rise in price level inflation, or are they counting money supply growth outpacing production growth as inflation?
Posted by: Jacob Oost at Apr 26, 2008 4:50:57 AM
Jacob: The average bundle of goods purchased by the average rich person differs from the average bundle of goods purchased by the average poor person. Since inflation is defined as the price increase of the average bundle of purchased goods, the rich and poor will see different inflation rates.
Posted by: David Wright at Apr 26, 2008 5:30:03 AM
Yeah, but inflation is a monetary phenomenon.
Posted by: Jacob Oost at Apr 26, 2008 8:45:17 AM
Can I see what Tyler's response is to "The Two Income Trap"?
Posted by: michael vassar at Apr 26, 2008 9:56:56 AM
Goods account for a larger share of poor peoples purchases and consumption,
but this does not necessarily mean that the poor are more likely to but new goods than the rich.
If a family with $30,000 income spends 50% of their income on nondurable goods or $15,000 while a family with a $100,000 income spends 15% or $15,000 on nondurable goods it does not follow that the poor are more likely to buy new nondurable goods even though they spend more of their income on nondurable goods.
Posted by: spencer at Apr 26, 2008 9:57:54 AM
Another way to interpret the Broda and Romalis work -- protectionism hurts the poor more than the rich.
Posted by: Richard A. at Apr 26, 2008 11:37:28 AM
@Jacob: What a way to misinterprete Friedman. Monetary policy has 'long and variable lags', in other words one percent of extra money will not raise the price of every good with one percent. A lot of U.S. money has been sucked up by the housing market, and I would say that the poor spend a smaller proportion of their income on housing then the rich. Therefore general inflation - even with it being 'a monetary phenomenon' - can have widely differing impacts on different groups.
Posted by: JSK at Apr 26, 2008 12:26:42 PM
The quality of the goods apparently does not count in this model. Lead poisoning takes a while to show up so I guess it works.
Steve
Posted by: steve at Apr 26, 2008 12:35:39 PM
For the poorest I had no doubt this is the case. They don't work in import competing industries so their wages are not affected by trade. Those who work in import competing industries are generally better paid whether skilled or unskilled and suffer much more being pushed down into the working poor resulting in a diminishment of the middle class. If you like the third world, you will love the future.
Posted by: Lord at Apr 26, 2008 1:15:28 PM
This is a fascinating result but I have a very basic question.
Is it really possible that economic researchers haven't ever done simple longitudinal studies asking a random cross-section of the population
to keep track of their purchases/income over many years, after which the numbers are analyzed to determine rough estimates of things
like the disposable income levels and purchasing habits of poor people? (Not to mention income mobility, etc.)
Or if these studies have been done, why haven't such studies suggested this result before? Small sample size?
Posted by: mk at Apr 26, 2008 8:20:08 PM
Yeah, I can see that. The rich guy who buys a classsic, but old Rolls Royce for his collection is just the offset for the poor guy struggling to make the payments on a Plymouth Neon.
Posted by: Max at Apr 26, 2008 9:16:10 PM
DRDR, that's a very good point. When I talked to Broda in writing the post, he mentioned that the phenomenon you describe probably means that the effect the paper describes will likely soon end as China's exports become more sophisticated.
Posted by: zubin at Apr 26, 2008 11:13:59 PM
Tyler,
The article you quote (in Portfolio) on the Broda/Romalis results states one of their key conclusions incorrectly. The Portfolio article states, "Broda and Romalis construct inflation rates for different income groups and find that rates for the richest outpaced rates for the poorest by about 4 percent over the period." If that's all they found, the result would be ho-hum. What makes it "shout from the rooftops" is that the rates for the highest-income (not richest) outpaced rates for the lowest-income (not poorest) by 4 percentage points, not 4 percent.
Best,
David
Posted by: David R. Henderson at Apr 27, 2008 9:52:02 AM
Uh, JSK, what are you talking about? Are you sure you're responding to something *I* said? I didn't even bring up Friedman.
And I have no doubt that inflation affects the poor the hardest, as do pretty much all economic hardships, but I do not see how the rich and the poor can be experiencing different levels of inflation when they are both using the same money supply.
The only way I can figure it is that the velocity of money among the poor is higher than that of the rich, so goods and services that are bought primarily by the poor experience inflation, while those bought by the rich don't......but I'm not a real economist and I don't even know if that could work, or if that phenomenon really ought to be called inflation or not. I don't see how inflation can be contained to one income class. Surely it spreads?
Posted by: Jacob Oost at Apr 27, 2008 10:48:04 AM
So the poor gain from trade in terms of expenditure.
To say if trade is beneficial overall, don't we need to know how trade affects incomes too? - for example by combining this study with work that might be available on one of the following phenomena:
- What role did trade play in the observed widening inequality?
- Are the real wages of the poor adversely affected by trade?
- Can movements of people into or out of the lower income sections of society be attributed to trade?
Posted by: Miles at Apr 27, 2008 11:22:32 AM
Miles,
Read all of the excerpt, not just highlighted portions about inflation. The paper not only said the poor have experienced lower inflation, but also applied those numbers to typical inequality measurement. The paper's main point was to answer your first point.
As to your third question, that's something that's very hard to quantify. Certainly there have been textile or steel workers who've gone from 20 an hour to 10 an hour at Wal-mart. However, manufacturing jobs peaked in the 1950s and declined since mainly due to greater automation. You might be surprised to learn, for example, that the number of manufacturing jobs in China has declined considerably since 2000.
Posted by: MW at Apr 27, 2008 6:02:51 PM
Certainly there have been textile or steel workers who've gone from 20 an hour to 10 an hour at Wal-mart. However, manufacturing jobs peaked in the 1950s and declined since mainly due to greater automation.
If automation is the biggest factor we should have lots of new/renovated factories that employ few workers but have high output.
A casual drive thru Michigan or Ohio will disabuse any non-economist from that idea.
Factories moved to China because workers are cheap, not because their productivity is higher. Most often it is lower.
However in China manufactiring employment very well may be declining because of increase in productivity.
Posted by: mik at Apr 29, 2008 5:03:12 PM
