Am I really that predictable?
Here’s Alex Tabarrok:
I just returned from a trip to South Korea. Today, to prepare for the next trip, I took my jacket to the dry cleaners. Turning the pockets out, I discovered a substantial number of South Korean won. The transaction costs of exchanging the won for dollars are now very high. I will keep the won as souvenirs.
Question: What are the consequences of my decision for the South Korean economy? Answer in the style of a well-known economist. What would Scott Sumner say? (almost too easy!) What about Keynes? Krugman? Cowen? Prescott?
I presume the South Korean central bank is targeting some set of macro variables, probably including interest rates, inflation and exchange rates. If so, Alex’s currency hoarding will have no impact on Korea’s NGDP, as the central bank will simply accommodate the extra demand for currency, just as the Fed accommodated the surge in foreign demand for US currency during recent decades.
If the currency is held for 10 or 20 years, then returned to Korea, it will represent an interest free loan to the Korean people. If it is handed down though generations, and eventually lost or made obsolete by new currency, then Alex’s hoarding would be a gift to the Korean people.
I’d like to think that the other 4 economists would answer this the same way, as I can’t imagine any other plausible answer. But perhaps I’m missing something.
PS. In the preceding post Tyler Cowen discusses a study of the British labor market:
2008 and after: -8.5%
That measure of wage decline is from John van Reenen (pdf, useful powerpoints on UK productivity), citing Martin and Rowthorn (2012).
Now I am all for the UK trying ngdp targeting, or for that matter well-targeted fiscal policy, or both. I never favored their *tax increases*, often misleadingly labeled “austerity” for political reasons.
I would, however, like to get a handle on Keynesian thinking here and thus the questionnaire aspect of this post. In the traditional Keynesian story, stimulus lowers real wages through nominal reflation. Is that the Keynesian view here? If so, why do Keynesians believe that British real wages need to fall more than 8.5% Why did they need to fall 8.5% to begin with?
The data provided in the study are weekly wages including bonuses (remember the City of London has highly volatile bonuses.) But the Keynesian sticky-wage model (a model I don’t accept) predicts that hourly real wages will rise in a demand-side recession. Because hours worked per week might fall in a recession, there is no prediction for weekly real wages. I don’t have the hours worked data in front of me, but I’ve read that part time work has recently soared in the UK. Perhaps someone can find the data. Certainly the 8.5% figure is fishy, as inflation is little more than 8.5% since 2008, and I’m sure British hourly wages have risen substantially since 2008.
One other point. A few days ago Tyler pointed to data showing a huge drop in the US employment to population ratio. I suggested that this data was unreliable. But let’s suppose that Tyler is right, and the ratio does tell us something about labor market health. It’s worth noting that this ratio has been pretty stable in Britain, falling only one percentage point since early 2007. And the paper Tyler cites here refers to the surprising health of the British labor market (which presumably relates to the same ratio.) So although I wouldn’t make this argument, I suppose it could be argued that (if British real hourly wages did fall) it helped keep the British employment to population ratio at a relatively lofty level.
I’d argue something slightly different. The increase in British part time work helps explain both mysteries; why real weekly wages did poorly, and why the employment to population ratio did well. And I’d add that the study also shows an absolutely abysmal productivity decline in Britain, which is much different from the US. Is that real or simply an artifact of measurement techniques? I have no idea. But either way it helps explain the GDP numbers. Either GDP is higher than measured, or the productivity has fallen enough to explain the decline in hourly real wages (assuming they did decline.) Hence there’s no mystery there.
PPS. Has British oil production declined in recent years? This might also play a tiny role in the mystery, as oil production is not very labor intensive, especially at the margin. Ditto for high finance.
PPPS. This is not to say that Tyler is wrong about Britain having structural issues. I’ve often made the argument that supply-side problems are part of the British RGDP mystery.
Tags:
22. October 2012 at 07:10
“The increase in British part time work helps explain both mysteries; why real weekly wages did poorly,”
Something I would add. In the UK part time hourly wage rates are significantly below full time hourly pay rates. As much as 30% for both men and women.
Not sure what the effect on your calculations there would be but it is a truth of that labour market.
22. October 2012 at 07:33
“If the currency is held for 10 or 20 years, then returned to Korea, it will represent an interest free loan to the Korean people.”
This should be obvious. The argument posed by those opposed is that extending interest free loans to a country is a bad thing for that country.
It is the same argument being posed by those individuals who believe that short term US interest rates below the inflation rate in the US are a bad thing, because it encourages the US govt. to borrow for consumption.
There are only really two arguments to support this, and both require a bit of augmentation to classical macroeconomics:
1) intergenerational equity (see Rowe’s recent example)
2) the ability to borrow relaxes budget constraints, which are the only thing that keeps political distortions to the budgeting process in check – it’s a principal agent argument
These arguments are poorly expressed, however
22. October 2012 at 08:02
ssumner:
I think it is important to emphasize how methodological collectivism can lead one astray in economic reasoning.
The argument here is that “the Korean people” are benefited by an influx of currency. The unstated premise is that “the Korean people” is a single conceptual entity, and is capable of deriving gains and losses due to various actions such as introducing a quantity of currency.
However, upon closer inspection, the detail of which only methodological individualism can offer, then what actually happens with the currency influx is that some Korean individual or individuals (depending on if the Won cash is split among more than one individual) will receive the new currency first, while every other individual in Korea will of course not receive the new currency first.
The effect of this currency influx will be to raise the nominal incomes of those initial receivers, while at the same time devalue the currency incomes of every other Korean. The initial receivers will be able to increase their nominal demand for goods, which is a gain to them, but it comes at the cost of every other Korean having a devalued income. They will be able to purchase less, because the demand for and prices of goods rises by more than their incomes.
Thus, contrary to the claim that a currency influx is a “gift to the Korean people”, in reality, a currency influx imbues the Koreans with no general benefits whatever. Some Koreans will gain, while other Koreans will lose.
This principle, that there is no general benefit to currency influxes, can be generalized to include all forms of money as a medium of exchange. If the Korean central bank increases bank reserves for example, this does not provide Koreans with general gains, but rather some Koreans will gain while other Koreans will lose, for the same reasons as above.
———————
A common objection by monetary central planners to the above is to say “OK, if an increase in the supply of money does not confer any general benefits, wouldn’t it be a waste of resources to mine gold in a gold standard, and isn’t it a good thing that the government enforces a non-gold monetary regime?”
This argument may sound plausible those who hold no principled objections to state violence in otherwise peaceful monetary affairs, but it overlooks a crucial point: Gold is not only money, but also, inevitably, a commodity. An increase in the supply of gold may not confer any monetary benefit, but it does confer a non-monetary benefit (e.g. electronics, jewelry, dental work, and in production processes such as industrial tools and equipment). Gold mining, therefore, is not a waste at all.
With fiat paper imposed by a state however, there is no general non-monetary benefits conferred by inflation. For what is happening? The state is defacing cotton and linen with ink (federal reserve notes), they are increasing the digital account numbers of favored individuals (reserve balances). Unless there are individuals who actually value the cotton, linen, ink and electrons in the form they take as money, for some other use than money (which can be safely assumed as utterly absent), then we can conclude that there is no general benefits whatever from fiat inflation. Not monetary, and not as a commodity either.
This is part of the reason why money is best left to the market. Not only will it help misguided economists how money works (by providing empirical minded pundits with the only thing that can help them understand anything: historical data), but it will also match up individual relative marginal utility of the money commodity, with the relative resource allocation to its production, which will prevent undue inflation and deflation, undue meaning too much or too little money relative to what the market process would have generated, in accordance with market signals denominated in the same commodity itself. In other words, only if money production is absorbed into the market process itself, can the commodity that is money serve as a tool for economic calculation in the market, rather than a tool of the state that hampers it.
22. October 2012 at 08:10
From David Smith’s column in the most recent Sunday Times (http://www.economicsuk.com/blog/001762.html):
“Though most sectors of the [British] economy are suffering from weaker productivity growth, two stand out. One is North Sea oil and gas, where output per hour has dropped more than 40% in five years.
“The other is financial services, where productivity was growing by more than 4% a year, but in the past three years has been falling nearly 3% annually, as its output has plunged. Just these two sectors provide much of the explanation for the very weak productivity numbers.”
22. October 2012 at 08:48
Wait, so what’s the difference between your model of wages and the Keynesian one?
22. October 2012 at 08:52
“If the currency is held for 10 or 20 years, then returned to Korea, it will represent an interest free loan to the Korean people.”
It doesn’t have to be held for 10-20 years, nor returned to Korea to be an interest free loan to the Koreans.
However, the ammount of money that leaks out of Korea in tourists pockets is trival compared to what goes on on the institutional side. Korea has unique laws which make it illegal for foreigners to hold KRW. The market has created the NDF to circumvent the currency controls. What are the implications of these currency controls and of the NDF market?
22. October 2012 at 09:06
Scott,
This is one of those posts where your postscript is far better than the main post.
Today Lars Christensen did a post on bubble fears in China, Poland and Sweden and Marcus Nunes did a post on Latvia. And then you did a nice postscript on the UK. So it must be cross country comparison day in Market Monetarist land.
However, whereas your postscript on the UK seems spot on, I felt compelled to critically comment on the other two posts. Lars implied Poland has a property bubble when in fact real estate prices have been falling for over four years. Marcus suggested that Latvia had nearly filled its NGDP hole but failed to note that its trend growth in NGDP was over 15% a year in the decade through 2008Q2.
In short I’ve been making the rounds criticizing MM posts, but there’s no need here. Good job.
P.S. I wonder if Britmouse has anything to add.
22. October 2012 at 09:11
Here’s a nice little article that gets to the predictability aspect of you Scott:
http://www.politico.com/news/stories/1012/82677.html
Its only where you are blind, that you are unpredictable. It’s also the only place you are weak.
22. October 2012 at 09:49
Tim, That sounds right.
Statsguy. Maybe, but that’s deeper than I wanted to look.
Sean, That’s worth a post.
Saturos. I think Wages/NGDP per person is key, they look at W/P.
Doug, Sorry, I don’t know much about that aspect of Korean policy.
Mark, Thanks for the tip.
22. October 2012 at 10:03
So employment to population ratio is something worth talking about and explaining?
22. October 2012 at 14:53
John, Some people believe that.
22. October 2012 at 16:42
All the things you mention have been considered by others and discounted as explanations for the discrepancy mystery between GDP and employment. The increase in part time work is a partial explanation but does not account for how total hours worked is also up. Goldman and JP Morgan analysts have been saying since at least 2009, that they believed the Office of National Statistics was understating UK GDP. Just in Q2 this year the ONS was recording contracting output as employment increased in the same period.
There has either been a huge fall in productivity, GDP is not counted accurately or something unexplained is going on. Possibly huge forbearance by the banks on zombie firms because they do not want to recognise losses.
Falling North Sea output does help explain some of the sluggish GDP performance. Simon Ward had an interesting chart a few months ago tracking GDP ex oil and gas.
http://moneymovesmarkets.com/journal/2012/6/28/uk-2008-09-gdp-decline-revised-down.html
22. October 2012 at 21:54
In the comments to the Alex Tabarrok post:
23. October 2012 at 05:24
Scott, you’re saying Keynes thinks that if a miser hoards currency, he is providing a gift to everybody else?
(I agree with you–subject to the point about “the owners and friends of the S. Korean central bank” being the same as “the Korean people”)–but you really think a Keynesian looks benignly on currency hoards?
23. October 2012 at 06:35
ssumner:
Saturos. I think Wages/NGDP per person is key, they look at W/P.
What if a bunch of people left the work for to become capitalists who earn profits, which happens to lead to a reduction in wage payments relative to total incomes?
What justification is there for advocating for inflation here? What is the justification for devaluing my money balance and nominal income, as well as the money balances and nominal incomes of everyone else who will receive the new inflation last?
23. October 2012 at 07:45
The problem is only puzzling because economists tend to use macro data rather than company level data. Research based on individual companies rolled up at the sectoral level (assuming this is representative of the wider economy which isn’t too unreasonable)) highlights that it is the financial services sector that explains the breakdown between rising employment and output. The on-going restructuring of the sector in the UK accounts for labour hoarding (with employment levels falling) and the sector is still being hit by declining productivity due to a swathe of onerous regulations on top of substantially higher capital requirements. Other sectors are performing reasonably well, but as the UK financial services sector accounts for a third of GVA it impacts the macro data used.
http://www.creditcapitaladvisory.com/2012/09/25/why-uk-productivity-puzzle-seen-micro-foundations-particularly-puzzling/
23. October 2012 at 08:03
MF, it’s the other way around. A fall in Wages/NGDP would imply a need to contract NGDP in Sumner’s model.
I think the maximum number of entrepreneurial capitalists (anyone with a financial portfolio containing equity is a capitalist) is inherently limited, so I wouldn’t expect your effect to be particularly large – there’s not that much room for people to give up their day jobs becoming traders or entrepreneurs (“real” or financial). (This also has implications for equitable taxation, in my view…)
23. October 2012 at 08:24
Bob, see the comments on the original post:
23. October 2012 at 10:43
Saturos, I know lots of people agree with me that Krugman/Keynes would hate hoarding. I want Scott to tell me if he disagrees with us. (That’s what you’re getting at, right?)
23. October 2012 at 10:49
Mr. Sumner, would you care to provide a link to your comments on Britain’s RGDP mystery? I’m interested.
23. October 2012 at 11:42
Saturos:
MF, it’s the other way around. A fall in Wages/NGDP would imply a need to contract NGDP in Sumner’s model.
I thought that in Sumner’s model NGDP grows at 5% per year, period.
I had assumed that a transition from wages to profits would, temporarily, and initially, lead to a reduction in NGDP, because it takes time for any new output to be ready for sale, which would require inflation in the meantime to bring NGDP (and hence wages, as per another assumption of the model) back up.
But let’s say you’re right, and that if wages fell then deflation would be called for. Ok, then I will ask to consider a case of RISING wages, as peopl go from the ranks of capitalists to the ranks of wage earners. You would agree that would necessitate inflation, right? Well, the same question can be asked. What is the justification for devaluing my income if some random people go from being capitalists to wage earners?
23. October 2012 at 12:02
Bob, yes they would hate hoarding. But being Keynesians, they have difficulty distinguishing it from saving in general.
Krugman would agree with Scott’s second paragraph, and hopefully also the first. Keynes would possibly predict a rise in interest rates for the first (disagreeing with Scott), and agree with the second. But they would both hate hoarding – except Krugman would normally just call it saving, whilst Keynes would call it “thrift”.
MF, actually on second thoughts it’s simpler than that. The ratio is hourly wages : NGDP per capita. People dropping out of the workforce (for whatever reason) doesn’t change average hourly wages (much), and wouldn’t really affect policy.
But yes I think if hourly wages were to drop down from their sticky path for some reason, it would be better to contract NGDP (that’s why the optimal target is not NGDP but rather a wage index). But you’re right that NGDP target calls for holding NGDP steady in that case and accepting some slight suboptimal inflation (assuming the jump in wages was small). But there is really no reason to expect average nominal hourly wages to jump off their trend path for no apparent reason, whilst NGDP is and is expected to be stable. Unless, as mentioned, there is mass uneven exodus of workers from the workforce and the hourly wage distribution – that would shift the mean hourly wage. Or long gradual change could change the trend path – but Scott has said it’s OK to make one-off changes to the NGDP level path at long intervals in response to predictable things like demographic change.
23. October 2012 at 12:06
Actually, the target should really be NGDP per worker, shouldn’t it, not NGDP per capita. So policy should tighten when people leave the workforce.
23. October 2012 at 13:09
Actually, the target should really be NGDP per worker, shouldn’t it, not NGDP per capita. So policy should tighten when people leave the workforce.
Even if they “leave the workforce”….to become capitalists/entrepreneurs? Why should NGDP tighten up then?
23. October 2012 at 13:11
MF, actually on second thoughts it’s simpler than that. The ratio is hourly wages : NGDP per capita. People dropping out of the workforce (for whatever reason) doesn’t change average hourly wages (much), and wouldn’t really affect policy.
Not if you think about one individual, no. But what if over a period of 5 years, 10% or 15% leave the workforce, to become capitalists? Wouldn’t it be significant then? I don’t think it’s good enough to assume no significant changes. You have to cover the possible corners (IMO).
23. October 2012 at 13:37
The ONS don’t have an “official statistic” series for hourly wages, but they do have some stats which they publish as “less than usually reliable”.
Looking at period 2008Q1 to 2012Q2; absolute changes (not annualized):
Nominal median hourly wages for all employees: up 7.9%
CPI level: up 15.5%
actually the ONS series for hourly wages had
23. October 2012 at 14:59
…impromptu suspense.
23. October 2012 at 19:04
Bob, Keynes would presumably have agreed with me if it was a fiat money example not at the zero bound. But I can’t be sure.
Britmouse, That CPI number is striking. Isn’t it the case that the VAT was at the same level in 2008:1 and 2012:2? I.e., the rate was cut after 2008:1 and raised before 2012:2? So what’s going on with the UK CPI?
In the past you’ve been skeptical of the supply-side take on Britain, but those CPI numbers are almost 4% per year. The US CPI is up only about 6% since mid-2008. Britain really is very different from the US.
Wage growth is a bit less than I would expect, maybe partly due to the use of median wages, not mean. But it looks like real wages fell more than I thought, although less than 8.5%.
You’ve still got a rising W/NGDP ratio.
24. October 2012 at 02:55
Couldn’t I point to the US CPI in late 2007/early 2008 as evidence of supply-side problems just as easily? The US CPI was several points higher than the UK CPI for quite a while.
Five years to September 2012, annualized growth rates:
US CPI: 2.1%
UK CPI at constant taxes: 2.7%
The major difference is that the US had sharp deflation in 2009 which the UK completely avoided, I’d say thanks to Sterling devaluation feeding through fast to consumer prices. That’s a supply-side shock but it was a fast recovery, UK CPI-CT tracks the US CPI very closely from 2010 onwards, the GDP deflator is even a bit lower for the UK if you strip out the VAT effect.