How to think about France
About once a year Paul Krugman does a post discussing the difference between the US and French economies. Here is Krugman in 2011:
So, here are some [2008] ratios of France to the United States:
GDP per capita: 0.731
GDP per hour worked: 0.988
Employment as a share of population: 0.837
Hours per worker: 0.884
So French workers are roughly as productive as US workers. But fewer Frenchmen and women are working, and when they work, they work fewer hours.
Why are fewer Frenchmen working? As I’ve pointed out, during prime working years they’re as likely to work as Americans. But fewer young people work (in part because of more generous college aid); and, mainly, the French retire earlier. The latter is arguably the result of misguided policies: Mitterand made early retirement alarmingly attractive. But it’s not a problem of weak productivity or mass unemployment.
And why do the French work shorter hours? Probably for the most part because of government policies mandating vacation time.
The bottom line is that France is a society with the same level of technology and productivity as the US, but one that has made different choices about retirement and leisure. Vive la difference!
I’m an academic like Krugman, and not surprisingly I also prefer more leisure time to more GDP. But one “difference” Krugman does not mention is unemployment, which has generally been higher in France than the US over the past 30 years, often by a wide margin (and hence it’s not just the euro problem.)
A few months ago Krugman noticed that France isn’t just poorer than the US, it’s also growing more slowly. That made him slightly less optimistic about France (as compared to the Ostry et al paper he discusses in this post):
Once you delve into this low labor input, it starts to look like the result of some very specific policies rather than redistribution in general: a pension system that encourages early retirement, regulations that give the French shorter hours and much more vacation time than we get.
Overall, I am still mostly persuaded by the Ostry et al work, but I think we need to acknowledge that it’s not quite as slam-dunky as liberals might like.
And now he has a new post emphasizing the fact that France actually has a higher employment rate than the US among prime age workers (25 to 54.) He seems to think that undercuts supply–side position that Europe is depressed due to high taxes and benefits:
The truth is that European-style welfare states have proved more resilient, more successful at job creation, than is allowed for in America’s prevailing economic philosophy.
I mostly agree with Krugman on the facts, but I see them as strongly confirming the Prescott/Mulligan view of the world. Perhaps there’s a framing affect problem here, as Prescott once compared France to Depression-era America, mostly on the basis of the very low level of hours worked. But let’s get past that misleading analogy, and think about what the supply-side model actual predicts. Keep in mind that France has a wide range of policies that reduce aggregate supply:
1. High taxes and benefits, which create high MTRs.
2. High minimum wages and restrictions on firing workers.
What should we expect from these “bad” supply-side policies? I’d say we should expect less work effort at almost every single margin. Earlier retirements, more students staying longer in college, longer vacations, and a higher unemployment rate. And that’s exactly what we see. We might also expect lower productivity. After all, those high tax rates should discourage capital formation, which would result in lower productivity. And yet the figures Krugman cites suggest that productivity is only 1.2% lower than in the US. That doesn’t seem so bad.
But here’s what Krugman misses. His other argument, that prime age employment in France is pretty high, strongly suggests that the productivity numbers are distorted by composition bias. Think about the people who are not employed in France, but who would be employed in America. The rate of employment among French immigrants from North Africa is presumably lower than among Latino immigrants in the US (which seems like the most comparable ethnic group.) The young in France are less like to work, as are the elderly. Given that French employment is much more heavily concentrated in prime age workers, I’d expect higher hourly productivity in France. I wonder how productivity varies controlling for age and ethnicity?
That’s not to say the supply-side model explains everything. I’m not sure why the rate of employment for prime age workers is lower in the US than France, but I’d guess it reflects the fact that the French model does have some advantages over the US model. Just off the top of my head I could imagine these might include a better K-12 system, less incarceration for drug “crimes,” fewer people on disability, etc.
But overall France conforms to the predictions of supply-side economics. The French are smart enough and efficient enough to spread the cost of not working far more effectively than did the US in the 1930s. The psychic pain of being unemployed is much less if you are a young person in school, or someone who retires early, or are spending the month of August in Provence. But their regime is not cost-free. They do have much higher unemployment, and that does impose psychic costs. Liberals constantly remind us (correctly) that unemployment is a devastating problem. And their per capita GDP is only 73% of US levels (in 2008, surely even less today.) And that percentage is gradually declining. Migrants from China and India tend to prefer the US model. So while I think France has a lot of good qualities such as top-notch infrastructure, and indeed arguably has close to the highest quality of life in the world, there is always room for improvement. And the labor market is one area where France falls short, just as the supply-side model predicts. Even modest reforms (say along German lines, not American lines) would deliver important gains. Instead, Germany is moving in France’s direction.
And finally, I don’t think it makes much sense to compare a medium size homogenous country like France to a large heterogeneous country like the US. Krugman has another post pointing out that the European welfare state model doesn’t work as well in Italy. His post is entitled “What’s the Matter with Italy?” and ends with the following:
I’m not going to answer this; truly, I don’t know. But it’s important.
He’s addressing the terrible Italian productivity numbers since 2000. I also don’t know why Italy has declined so sharply in recent years. But there is one thing we do know about Italy, the low level of per capita GDP (compared to France) is almost entirely due to an absolutely horrendous performance in southern Italy, where roughly 1/3 of all Italians live. And there is a lot of circumstantial evidence that at least part of the difference between southern and northern Italy is cultural. Not “cultural” in the sense that some people use the term (a code word for lazy.) After all, southern Italians do quite well in America. Rather cultural in the sense that everyone’s hobbled by a culture of corruption than no single Sicilian or Neapolitan is in a position to change.
So I don’t believe you can think about France vs. the US without also thinking about southern Italy. That’s part of the rich mosaic that is the Welfare States of Europe (the WSE), just as South Dakota Indian reservations, and McAllen Texas, and rural Mississippi are part of the United States of America (USA), the world’s richest big economy. It would probably make more sense to compare France to an above average region of the US, such as New England or the mid-Atlantic states. In that case the problem with the French labor market model would be easier to see. The problem is not that the French model has “failed,” as some conservatives suggest. The problem is that France has not achieved Swiss or New Hampshire levels of success, despite having the human capital and cultural traditions needed to do so. They should aim even higher.
HT: Edward
PS. In 2012 the World Bank has French GDP/person (PPP) at 71.1% of US levels. In 2013 the IMF has French GDP per capita at 67.4%, while the CIA has 67.6%. I suppose you could argue the drop since 2008 is cyclical, except no one is predicting French GDP/person to rise faster than US GDP/person over the next few years. It looks permanent to me.
As a point of comparison, African American housholds make 65.3% of the US average, and Hispanics make 76.5% of the average. Is that comparison misleading? That’s an understatement; it’s utterly and completely misleading. Apples and oranges. I simply provide the data for all the fools who think income data tells us something useful about issues like economic inequality.
PPS. I just noticed that Tyler Cowen has an excellent post on this topic.
Tags:
28. May 2014 at 06:54
My high school age daughter just spent a couple of weeks at a similar school to her own (Princeton, private) in Paris.
Quick take: Education there is still very top down, very 1970s. Nary an iPad to be seen (which is now standard in private schools in Princeton).
Students in Paris smoke a lot, drink a lot. She had a good time.
As for Italy, well, if Krugman doesn’t have an idea, I do. Also, I think Berlusconi bears much blame for the weak economic performance on the last decade in Italy.
28. May 2014 at 07:52
To be fair to France, they are laboring under the ECB’s anorexic monetary diet and most of their trading partners too. The long-term effects of too-tight money are corrosive and cumulative; see Japan.
Imagine if for the past 10 years the ECB has been tenaciously and aggressively growth-oriented…
28. May 2014 at 08:04
I was just thinking about compositional effects, too.
http://idiosyncraticwhisk.blogspot.com/2014/05/revisiting-minimum-wage-workers-as.html
In 1985, when the minimum wage was 36% of the average wage, about 5% of workers were at 36% of the average. Now that MW is at 28% of the average, more than 10% of workers are at 36% or less than the average. As someone who thinks MW kills jobs, I look at that and say, “Alright! 5% of the working population now has a job that was unavailable before!” But, if you don’t think MW kills jobs or if you don’t account for it, it looks like the economy of the past 30 years has produced a lot of very low paying jobs. If this is your point of view, then your reaction might reasonably be, “This economy is set up against the poor. It’s only producing low quality jobs. Let’s raise the minimum wage to help low wage workers get higher wages.”
This compositional effect will always make economies that produce opportunities for low skilled workers look worse.
28. May 2014 at 08:23
“French workers are roughly as productive as US workers. But fewer Frenchmen and women are working, and when they work, they work fewer hours”
This is not so far from “incarceration rises despite lower crime rates” reasoning. Let me rephrase it: “Productivity is higher and low-productivity workers are kept out of the labour force.”
28. May 2014 at 08:57
There are a couple of escape valves in the French system that probably help to boost prime age employment. France has a higher fraction of its workers in temporary/contract jobs than the US, which isn’t really a big problem because the welfare system there is much better. The heaviest labor regulations also don’t kick in until a firm hires 50 employees or more, so you get more smaller companies.
Like Great Britain*, that’s probably being dragged down by less productive rural areas. The GDP per capita PPP in Paris, for example, is much higher than the national level.
* GDP per capita in London and New York City (both international financial capitals) is comparable while Great Britain’s GDP per capita is lower than the US.
28. May 2014 at 09:50
Kevin –
Have you–or can you–estimate the impact of no minimum wage on US employment?
S.
28. May 2014 at 10:40
but one that has made different choices about retirement and leisure. Vive la difference!
—
I’m an academic like Krugman, and not surprisingly I also prefer more leisure time to more GDP.
More accurately, the French system prevents people from choosing less leisure, while the US systems allows it.
France actually has a higher employment rate than the US among prime age workers (25 to 54.)
What does the rest of the age distribution look like?
28. May 2014 at 11:11
TallDave,
Correct. That’s connected with the fallacy that Krugman makes when he says-
“The bottom line is that France is a society with the same level of technology and productivity as the US, but one that has made different choices about retirement and leisure.”
If French society had made that choice, then they wouldn’t need state enforcement of that “choice”.
28. May 2014 at 11:19
Steven, 2006 was pretty close to a non-operative mw. Even now disemployment is probably fairly minimal. A broad estimate I would use is about a 1% decrease in employment for each $1 increase.
28. May 2014 at 12:27
Germany is moving in France’s direction. Protecting incumbents is pretty well the default position on regulation in general and labour market regulation in particular. Perhaps even an equilibrium position. In a manner of speaking.
28. May 2014 at 12:41
More accurately, the French system prevents people from choosing less leisure, while the US systems allows it.
Hedonic treadmill.
Let’s face it – Westerners have their material needs met. And then some.
28. May 2014 at 13:08
Kevin –
OK, so let’s say that it’s 1% per dollar. Then reducing the minimum wage to $8 is worth 1.4 million jobs. Yes? Maybe?
There are currently 258,000 blacks of both sexes, aged 16-19 who are unemployed (37% unemployment rate). Now, what percent of those 1.4 m jobs would black teenagers get? 9%, maybe?
Then that’s 126,000 jobs. It would cut black teenage unemployment by half.
Is that what we’re saying?
http://www.bls.gov/news.release/empsit.t02.htm
28. May 2014 at 13:39
When a Spanish economist, ensconced at Columbia, reviews Piketty, hilarity ensues;
http://salaimartin.com/randomthoughts/item/720-piketty-y-capital-en-el-siglo-xxi.html
Cleaning up the Google translation a little;
————quote————-
If the most famous thesis of [Thomas] Piketty is true and the rate of return on capital is higher than the growth rate of the economy (r > g), it follows that we should have a capitalized social security system . I.e., with the current system of PAYGO in which young workers pay taxes on their wages to supply the pensions of retirees, the rate of return they get is the growth rate of wages, ie, ‘g’. If, on the contrary, we take the money wages of those younger workers and we invested it in capital would get a return ‘r’.
If, as Piketty says, “r > g “, retirees could enjoy a much higher pension if the pension system was fully funded! Which is what I argued to the union leader and former leader of ICV, Joan Coscubiela in a debate on the future of work, we did on TV3 [in Spain] . He said the return on capital was much higher than the growth rate of wages and that meant that capital “ate” an increasing share of GDP. I warned him to be careful of that argument because the immediate implication was that the model of a fully funded pension is superior to PAYGO. I got the impression that Coscubiela did not understand the contradiction of his own argument, but Thomas Piketty does understand that.
He devotes an entire section of his book to discussing pensions. Recognizing the validity of the argument, Piketty explains that now it is too late to change the system because at the time of the transition we would have a generation that has to pay twice [My note; that’s not a valid argument, as Milton Friedman has demonstrated]. But that argument is not valid for emerging countries that do not yet have a pension system. I wonder if Piketty (and Coscubiela) advocate the introduction of a capitalized private pension scheme for poor countries that do not have social security.
In fact, the argument is not valid for the rich countries: if truly the return of capital is so superior to wage growth, you could take some of the excess return, save it and grow it until you have enough money to make the transition . Why not defend that, Piketty!
Then there is a second reason for not wanting a funded [capitalized] system: the return on capital is higher, true, but it is also vastly more volatile and uncertain! That is, after writing an entire book on the great bargain it represents for the rich to have capital, at the end of the book he confesses that some of the higher return is compensation for taking higher risk. A risk that Piketty does not want workers to take. Curiously, Piketty does not analyze the risk assumed by capitalists throughout the book. Piketty speaks of the rate of return of capital as an overpayment to a class of citizens who do little more than suck the blood of the workers. At the moment of truth, however, Piketty confesses that, at least in part, that “r” rewards risk taking by those who invest. It’s one thing simply to get a return for exploiting your fellow citizens, and quite another to reap a reward for taking a risk that the rest of society (and Piketty most of all!) is not willing to take.
This lack of rigor in analyzing the relationship between return and risk of capital is one of the major shortcomings of the book.
————–endquote————
Sala y Martin’s lengthy review came to my attention thanks to Jim Glass, this morning.
28. May 2014 at 14:32
Steven,
The national MW is $7.25 now. Below $7, I think the disemployment effects would be muted.
While it probably does have a sharper effect on employment among young black workers, it’s tough to say how much with much specificity.
I usually talk about it in terms of changes in employment, because many MW jobs are marginal, so much of disemployment comes through movements out of the labor force.
28. May 2014 at 14:50
Yglesias points out a key metric where France looks great physical fitness!
http://www.vox.com/2014/5/28/5756110/so-much-for-the-mediterranean-diet-greece-italy-have-the-worlds-most
P.S.: Does tight money help explain why boys are relatively overweight in Greece, Italy and Spain?
28. May 2014 at 17:31
Thanks, Kevin.
28. May 2014 at 18:40
Everyone, Yes, I agree with most of the comments. It’s certainly the case that the French choose much less leisure when they were freer to choose (i.e. in the 1960s.)
29. May 2014 at 03:17
Patrick: since Marx’s economic analysis is basically founded on ignoring risk, there may have been a contagion effect. Or, perhaps, such ignoring is just necessary to get specific conclusions.
29. May 2014 at 06:04
“He seems to think that undercuts supply-side position that Europe is depressed due to high taxes and benefits:…”
“…I mostly agree with Krugman on the facts, but I see them as strongly confirming the Prescott/Mulligan view of the world. Perhaps there’s a framing affect problem here, as Prescott once compared France to Depression-era America, mostly on the basis of the very low level of hours worked. But let’s get past that misleading analogy, and think about what the supply-side model actual predicts. Keep in mind that France has a wide range of policies that reduce aggregate supply:
1. High taxes and benefits, which create high MTRs…”
“…What should we expect from these “bad” supply-side policies? I’d say we should expect less work effort at almost every single margin. Earlier retirements, more students staying longer in college, longer vacations, and a higher unemployment rate. And that’s exactly what we see. We might also expect lower productivity. After all, those high tax rates should discourage capital formation, which would result in lower productivity. And yet the figures Krugman cites suggest that productivity is only 1.2% lower than in the US. That doesn’t seem so bad…”
I’d like to address the tax aspect of this, but first a word about the productivity numbers.
They came from the OECD and they’ve discontinued the series. The Conference Board (TCB) constructs annual productivity (GDP per hour worked) estimates for a far wider sample spanning a longer period of time (and it’s still ongoing) and what’s more it obtains its source data for the OECD countries from the very same survey that the OECD obtained its source data.
TCB’s estimates of relative French productivity were always a little lower than the OECD’s, but what they show is that French productivity reached a relative high of 98.5% of US productivity in 1995. It fell to 94.1% of US productivity by 2006 and this has been attributed to the IT boom in the US. But then it fell off a cliff to 88.5% of US productivity by 2010 and as of 2013 it was down to 88.1% of US productivity.
Now for the taxes.
France has a less than ideal tax system, but then what country doesn’t? I’d argue that US tax system is actually worse. The only virtue of the US tax system is its relatively low *level* as measured in terms of tax revenue to GDP. In terms of tax *structure* the US system is one of the worst.
One measure of tax structure is to look at the implicit tax rates (ITRs) by type: consumption, capital and labor. Implicit tax rates are the amount of tax revenue divided by the tax base and so are essentially the average tax rate. Unfortunately recent comparable ITRs for the US are not available, but fortunately they don’t tend to change at too fast a rate. Here’s the European Commission’s estimates for comparable measures through 2001:
http://ec.europa.eu/economy_finance/publications/publication11064_en.pdf
The consumption ITR (CITR) can be found on page 72, the capital ITR (KITN) on page 78 and the labor ITR (LITR) on page 70. As of 2001 the US and France had a CITR of 23.9% and 9.1%, a KITN of 36.9% and 39.9% and a LITR of 24.0% and 42.8%. So in short, France taxes consumption nearly three times more heavily, labor nearly twice and heavily and capital only slightly more heavily than the US.
Another measure of tax structure is to look at proportion of revenue derived from various sources. Unfortunately the European Commission has no comparable numbers for the US but its possible to come up with some rough estimates using OECD figures. The most recent complete numbers I have are for 2011. What I find is that total tax revenue is equal to 24.0% and 44.1% of GDP in the US and France respectively. So the level of taxation is far higher in France, but the literature on tax level and economic growth is rather inconclusive. This is one reason why the research has moved on to tax structure.
Consumption tax revenue is equal to 4.4% and 12.1% of GDP in the US and France respectively. Corporate and property tax revenue is equal to 5.5% and 6.0% of GDP in the US and France respectively. Personal tax revenue (payroll, social security and personal income taxes) is equal to 14.1% and 26.0% of GDP in the US and France respectively. It’s not a perfect correspondence since personal income taxes includes capital and property taxes (dividends, interest, rent and capital gains) but personal tax revenue largely corresponds to labor taxes. Almost exactly the same percentage of tax revenue (59%) is obtained from labor taxes by this crude measure. In contrast corporate and property taxes, which very roughly corresponds to capital taxes, represents 23% of tax revenue in the US and 14% of tax revenue in France. Consumption taxes constitute 18% and 27% of tax revenue in the US and France respectively.
The literature on tax structure and economic growth generally suggests that capital taxes are the worst and consumption taxes are the best with labor taxes in between. Thus the US tax system is clearly worse than the French system in that, by any measure, it undertaxes consumption, and it is also worse in that it is excessively dependent on capital taxation as a source of tax revenue.
Now, since we’re talking about quantity of labor supplied, I think should say a word or two about that.
One can talk about hours worked per worker and the employment share as Krugman does, or one can simply reduce it to a single number as the number of hours worked per working age adult (15-64). Total hours worked can be obtained from TCB and the working age population can be obtained from AMECO. Here are the numbers for 38 nations in 2013:
1.Luxembourg 1552
2.Japan 1426
3.Switzerland 1422
4.Korea 1400
5.Mexico 1379
6.Iceland 1350
7.New Zealand 1349
8.Australia 1314
9.Estonia 1292
10.Canada 1284
11.Latvia 1270
12.Portugal 1258
13.Czech Republic 1251
14.Sweden 1219
15.Austria 1207
16.Romania 1206
17.Lithuania 1204
18.Finland 1194
19.United Kingdom 1178
20.United States 1176
21.Poland 1154
22.Hungary 1154
23.Bulgaria 1153
24.Norway 1145
25.Cyprus 1121
26.Greece 1112
27.Italy 1107
28.Ireland 1100
29.Malta 1098
30.Denmark 1087
31.Netherlands 1078
32.Germany 1066
33.Slovakia 1008
34.Slovenia 1006
35.Belgium 977
36.France 951
37.Turkey 924
38.Spain 922
Yes, the French work very few hours. But the US is below the average and the median for this group. In particular I want to call your attention to Austria, Finland and Sweden. All three nations have higher LITRs than France, even according to 2011 European Commission data. All three nations also derive a higher proportion of their total tax revenue from labor taxes according to 2011 European Commission data. And for what it’s worth all three have higher top personal income tax rates (2013) than the US.
http://ec.europa.eu/taxation_customs/resources/documents/taxation/gen_info/economic_analysis/tax_structures/2013/report.pdf
If you take the 27 EU nations on the list and add in Norway and regress the hours worked on the 2011 LITR from the European Commission the R-squared value is 1.3%. If you take the hours worked and regress it on the top personal income tax rate from the European Commission the R-squared value rises to a still statistically insignificant 7.6% but the slope is negative.
In short there is absolutely no relationship between hours worked and labor taxes. There may be bad supply side policies restricting the number of hours worked in France, but labor taxes, and high MTRs on personal income taxes, is not one of them.
29. May 2014 at 07:47
Dear Scott,
Your analysis of the French situation is perfectly true. Krugman uses France whenever he needs to make his point but he does not know the country at all: if France was the paradise he describes with people choosing to work less to get more leisure, why do you have the far right at 25% ? The only explanation is that the current social model is rejected by the French. They want more jobs and more income, not more leisure. They want supply side policies (even if they would not put it that way given the terrible economics knowledge level there).
A French
29. May 2014 at 08:48
I’m thinking about copying and pasting this last comment from Mr Sadowski. Good stuff.
29. May 2014 at 08:53
A little more from My Breakfast with Xavier;
———-quote———-
…contrary to the claims of Piketty, the fact that r is greater than g implies neither the rich pass their savings on to their children, nor that wealth grows faster than GDP, or that rich dynasties are increasingly richer, or that social inequalities grow.
Imagine, for example, a world in which individuals work when young and retire when old, knowing that someday they will retire. So when young, they save money and invest. When they are old they use their savings (and the rate of return on their savings) to survive. It’s no longer a dollar inheritance to their children if they die with nothing.
The children do the same as their parents and thus, generation after generation too. All economists know that in this world of “overlapping generations” the rate of return on capital, r, may be higher, exceeding the growth rate, g. If the economy is dynamically efficient, then it is true that “r> g” and yet, no one gets an inheritance!
Unlike what Piketty says, logic does not dictate in any way, that “r> g” implies that inherited wealth grows faster than GDP, partly because inherited wealth can be exactly zero in worlds where “r> g”!
In the real world, of course, the rich do not consume everything they have and leave part of their wealth in inheritance to their children. It is also true that for many of them an important part of that wealth is spent on lavish parties, boats, airplanes, luxury travel or philanthropic actions such as Bill Gates or Warren Buffet.
Furthermore, unlike what happened in ancient times, where all the wealth went to a single heir, now the property of the rich is divided between many children (often from different marriages) so that from a very rich grandfather You can have very poor grandchildren. It is a well known saying that the grandfather created a fortune, his children extended the fortune, and the grandchildren squandered same.
The world could have “r> g” and, in turn, be filled with families whose grandparents create fortunes, the grown children and grandchildren destroy them. And contrary to what Piketty says, in that world there would not be more and more rich and powerful dynasties.
———-endquote———
29. May 2014 at 09:35
I am an academic in France.
Just one anecdote about taking “freely” more leisure.Every year my secretary takes a lot of days out, towards the end of the year. She doesn’t want to do that (hey, may be she enjoys the prof with whom she works…), but she is obliged to by law because she has to take her “RTT” -reduction in working time- before the end of the year; she would prefer to have extra days paid, but, no, not allowed by law!
29. May 2014 at 14:44
Exactly ! And to gain some flexibility, they have put in place “extra hours savings accounts” (compte épargne temps) in big companies: you are allowed to “save” x days of holidays per year and to use them later for different purposes (early retirement, special situations…). Of course, it costs a lot to corporations (you have to hire people to manage these accounts) but it is also a scheme which mostly benefits to insiders (people hired in big companies) who already have very satisfactory working conditions. What Krugman does not understand is that France is a country of insiders. When you are in the system, it may be the best country to live in the world: you enjoy long holydays, good wages, beautiful women, wine and cheese (pick your dream in the list). But when you are an outsider (and they are becoming more an more numerous), you are bound to remain where you are with almost no hope of getting up in social ladder. That is frustrating, and that´s why the far right is so high in the polls. France is very unequal and contrary to what Piketty and Krugman believes, redistribution and the welfare state are not going to solve that issue: you have to grow the pie before anything else.
29. May 2014 at 19:55
i haven’t thought this through yet but one thing that is true is that the minimum wage in france (known as the SMIC) is high. Currently $12 an hour. So by definition (almost) productivity has to be higher than $12/h or you are unemployed. One way the US could raise productivity would be to raise the minimum wage a lot. $15 say. Productivity would be much high (unfortunately employment would be higher and probably GDP would be lower but, hey, we are just comparing productivity).
29. May 2014 at 20:02
Oh, and when I was a manager in France for 5 years it was a pain in May. At that time (it sounds like the timetable might have changed based on the prof with a secretary comment) May was awful. First there are a zillion public holidays in May. Second, if you didn’t use your vacation by June 1st you lost it. So people would just stay home even if they had no reason to and veg out in front of the TV (this was pre internet in the very early 90s). So May in France was like Monday in Russia was reputed to be — very high “absenteeism” (people not at work for whatever reason).
30. May 2014 at 02:57
Mark, I mostly agree with your comments on taxes, although I would point out that consumption taxes ARE labor taxes, so I don’t find that distinction to be very meaningful.
I do have a question about the hours worked however. I have seen data on hours worked published in many different places, and in each case the results were very different from what you show. That doesn’t mean your data is wrong, but I’d like to see more confirmation.
In general, hours worked are much higher in East Asia than in the US, and taxes tend to be lower. In general hours worked are much lower in Europe than the US, and taxes are much higher. I realize your data doesn’t show hours worked being much lower in Europe, but again, that conflicts with all of the other studies that I have seen.
One possibility is that my studies are out of date, as hours worked in places like Germany have recently risen, whereas they have fallen sharply in the US. One question is whether this is merely a cyclical pattern, or if it’s structural.
Of course the flip side is that US productivity has recently done much better than European productivity, which is what your figures also show.
Everyone, Thanks for the comments about France, very good points.
30. May 2014 at 04:30
“Mark, I mostly agree with your comments on taxes, although I would point out that consumption taxes ARE labor taxes, so I don’t find that distinction to be very meaningful.”
Theoretically yes, but not in practice:
http://rlv.zcache.com/sure_it_works_in_practice_but_does_it_work_i_tshirt-r77d64a9794654db1adc7c6b16143371c_8natd_512.jpg
30. May 2014 at 04:31
“I have seen data on hours worked published in many different places, and in each case the results were very different from what you show. That doesn’t mean your data is wrong, but I’d like to see more confirmation.”
I’m not sure what you mean by “confirmation”.
Either the data is correct or it isn’t. I don’t think there is any question that the AMECO numbers are correct, and The Conference Board is the leading source for international productivity data.
Here’s the source data:
Country-Population-Hours
1.Luxembourg 375.0 582,032
2.Japan 79,272.8 113,079,919
3.Switzerland 5,477.4 7,789,428
4.Korea 36,676.7 51,343,038
5.Mexico 75,420.0 104,029,204
6.Iceland 215.5 290,853
7.New Zealand 2,955.5 3,987,731
8.Australia 15,314.9 20,124,770
9.Estonia 874.7 1,130,084
10.Canada 23,953.2 30,756,244
11.Latvia 1,348.7 1,712,388
12.Portugal 6,881.2 8,654,013
13.Czech Republic 7,147.3 8,938,220
14.Sweden 6,137.2 7,482,187
15.Austria 5,697.3 6,876,339
16.Romania 13,685.8 16,505,682
17.Lithuania 1,994.7 2,401,667
18.Finland 3,502.0 4,179,714
19.United Kingdom 41,834.3 49,280,822
20.United States 211,426.7 248,608,865
21.Poland 27,294.4 31,498,158
22.Hungary 6,773.0 7,815,046
23.Bulgaria 4,894.5 5,641,433
24.Norway 3,356.5 3,843,985
25.Cyprus 615.6 690,122
26.Greece 7,192.9 8,000,535
27.Italy 38,686.3 42,843,985
28.Ireland 3,012.5 3,313,394
29.Malta 287.2 315,277
30.Denmark 3,624.8 3,941,686
31.Netherlands 11,152.8 12,017,812
32.Germany 54,379.9 57,966,925
33.Slovakia 3,873.3 3,903,601
34.Slovenia 1,404.0 1,412,638
35.Belgium 7,308.5 7,142,033
36.France 41,882.9 39,823,898
37.Turkey 51,529.1 47,622,443
38.Spain 31,218.5 28,779,488
The working age population data comes from AMECO. Choose “AMECO online”, “Population and “15 to 64 years (NPAN)”
http://ec.europa.eu/economy_finance/db_indicators/ameco/index_en.htm
The hours worked data comes from The Conference Board.
Go to the bottom of the page (“Downloadable Files”). Download “Total Economy Database – Output, Labor, and Labor Productivity, 1950 – 2013”. Choose the tab labeled “Total hours worked”.
https://www.conference-board.org/data/economydatabase/
30. May 2014 at 05:11
“In general, hours worked are much higher in East Asia than in the US, and taxes tend to be lower. In general hours worked are much lower in Europe than the US, and taxes are much higher. I realize your data doesn’t show hours worked being much lower in Europe, but again, that conflicts with all of the other studies that I have seen.”
The only East Asian country for which I can find comparable labor implicit tax rates (LITRs) is Japan. (It is comparable to the US.) And the only East Asian countries for which I can find a reasonably detailed breakdown of tax revenue by tax type (courtesy of the OECD) is Japan and Korea. Both of these countries have low tax levels, but Japan does get over 60% of its total tax revenue from labor taxes, and its top personal income tax rate is 50.8% which puts it within the top ten countries in the OECD.
The only cross-sectional study I have ever seen that argued that there is a inverse relationship between the quantity of labor supplied and taxes is this one by Prescott and McGratten:
http://www.minneapolisfed.org/research/qr/qr2811.pdf
The empirical analysis involves a data set of only seven nations in two periods, and some of the residuals are enormous. I do not find it at all persuasive, especially given the other evidence that labor supply is relatively inelastic.
30. May 2014 at 06:56
“One possibility is that my studies are out of date, as hours worked in places like Germany have recently risen, whereas they have fallen sharply in the US. One question is whether this is merely a cyclical pattern, or if it’s structural.”
I think it’s a bit of both. US labor market participation peaked in 2000. And the recession has manifested itself in very different ways with respect to the labor market in the US relative to Europe.
To investigate how important a role the recession may have played I recalculated the hours worked per working age population numbers for the year 2007:
1.Luxembourg 1604
2.Latvia 1557
3.Korea 1542
4.Iceland 1504
5.Mexico 1422
6.Japan 1421
7.Portugal 1415
8.Switzerland 1411
9.Estonia 1408
10.New Zealand 1365
11.Cyprus 1348
12.Ireland 1318
13.Lithuania 1317
14.Greece 1303
15.Canada 1292
16.Australia 1291
17.Czech Republic 1245
18.United States 1244
19.Austria 1235
20.Romania 1218
21.Sweden 1214
22.Hungary 1207
23.Finland 1206
24.United Kingdom 1197
25.Italy 1190
26.Bulgaria 1182
27.Poland 1165
28.Norway 1160
29.Denmark 1158
30.Slovenia 1149
31.Spain 1102
32.Netherlands 1082
33.Malta 1060
34.Germany 1040
35.Slovakia 1009
36.Belgium 989
37.France 964
38.Turkey 850
Now the US is above the median but it is still below the average for the group. It is also above Austria, Finland and Sweden. (Incidentally the Czech Republic also has a higher LITR than France in both 2007 and 2011.)
Interestingly if you regress the hours worked for the 27 EU members plus Norway on the LITR, the R-squared is 10.3%, and since the p-value is 9.5%, it is just barely statistically significant at the 10% significance level. More importantly the sign of the slope is as expected (higher LITR, lower hours worked).
However, I still think other factors are far more important, as for example exemplified by the anecdotal evidence provided by the above French commenters.
31. May 2014 at 06:23
Mark, Thanks for the data. Here are a few replies:
1. The 2007 data is indeed closer to what I’ve seen elsewhere. So that explains part of the mystery.
2. In 2007 US hours worked were much higher than in Europe, due to the fact that all the big European countries had hours worked below US levels.
3. In 2007 hours worked in Greece and Portugal and Cyprus and Czech Rep. and the Baltics was high, as you’d expect in lower income countries. The other outliers included Iceland and Switzerland, two affluent European countries that (in 2007) had very unusually low top income tax rates. Don’t know about Ireland and Luxembourg, but aren’t they tax haves of some sort? All those countries combined have fewer people that Italy or France.
4. Labor supply is a tricky issue. The labor supply relevant for this issue is the income-compensated labor supply. Only the substitution effect is relevant; as I said the income effect pushes lower income countries to work more.
5. Hours worked in Singapore, Hong Kong and Taiwan are extremely high, and tax rates are extremely low. Add them in and I’ll bet you get a stronger result. Last time I looked Singapore hours worked were well over 2000. Top tax rate is around 15%. Japan has the highest tax rates of the 5 developed East Asian countries, and the lowest hours worked.
6. It’s not just cross sectional. In the 1960s the French had similar tax and spending rates to the US, and similar hours worked. Then tax rates soared well above US level, and work effort plunged. So it’s not “cultural differences.”
7. Regulations are endogenous. When tax rates are high people prefer to work less. Governments then cater to that demand (for a standard holiday schedule), by adding in many more days off, or a 35 hour work week.
8. I’m surprised Prescott’s the only one who’s looked at this, in a model with 4 or 5 variables it would be easy to get a decent fit for 35 developed countries, including 5 East Asian countries.
9. Yes, there are real world differences between consumption taxes and labor taxes, but they are very small. Again, a VAT is a labor tax, as I believe all VATs exclude capital costs. If the incidence is not identical to a payroll tax for technical reasons, it is extremely close. In any crude study (and all these studies are extremely crude) the two taxes should be viewed as essentially identical. Both payroll taxes and VATs discourage work.
10. Tax and benefit regimes are extremely complex, and no one number can capture the effects of the myriad of implicit MTRs.
Until someone comes up with a better hypothesis, I’m going to assume big government explains why the Europeans work less and small government explains why the East Asians work more. I do agree that regulations also play a role, and that it’s hard to disentangle all the various effects. It’s also hard to know what’s endogenous and what’s exogenous.
31. May 2014 at 06:26
Mark, Just to reiterate, any study must include consumption taxes as labor taxes, otherwise the results are simply not valid.
31. May 2014 at 09:43
Scott,
“In 2007 hours worked in Greece and Portugal and Cyprus and Czech Rep. and the Baltics was high, as you’d expect in lower income countries.”
True, but you also have countries such as Bulgaria, Malta and Slovakia. All three are relatively low income countries with a low number of hours worked per working age population in 2007. All three also have low LITRs, low top personal income tax rates and are less dependent than average on labor taxes as a source of revenue.
“The other outliers included Iceland and Switzerland, two affluent European countries that (in 2007) had very unusually low top income tax rates.”
Yes, but Iceland has a high overall tax load, equal to 40.6% of GDP in 2007 and Switzerland is unusually dependent on labor taxes with personal income taxes, social security taxes and payroll taxes equal to over 55% of total tax revenue in 2007.
“Don’t know about Ireland and Luxembourg, but aren’t they tax have[n]s of some sort?”
Both are classified as “tax havens” but this isn’t because of their low labor taxes. Ireland’s top corporate tax rate is only 12.5% which is lower than all of the EU-27 with the exception of only Bulgaria and Cyprus. Luxembourg’s top corporate income tax rate isn’t really that low (29.2%), but similar to the Netherlands it facilitates “conduit structures”
which make it possible to channel financial flows through in order to be charged lower taxes in other countries.
“I’m surprised Prescott’s the only one who’s looked at this, in a model with 4 or 5 variables it would be easy to get a decent fit for 35 developed countries, including 5 East Asian countries.”
I didn’t say Prescott is the only one who’s looked at this, I said he’s the only one I knew of who who found an inverse relationship between taxes and the quantity of labor supplied in a cross sectional study.
For example, here’s a 2005 study by Alesina, Glaeser and Sacerdote that finds that when you include union coverage rates, employment protection, and country and period fixed effects for 14 countries, the coefficient on taxes is not only statistically insignificant, it is the opposite sign of what is expected (Table 1.10, Column 4, on page 36):
http://www.nber.org/chapters/c0073.pdf
I looked for papers other than Prescott’s, and the only one I can find that has results similar to Prescott’s is this one by Ohanian, Raffo and Rogerson (2008):
http://eml.berkeley.edu/~saez/course/Ohanianetal(2008).pdf
But when Jon Bakija more or less replicated the work of Ohanian et al, and then included period fixed effects, the slope coefficient on taxes fell by nearly a factor of four (Table 2, Columns 2 and 4, on Page 8):
http://web.williams.edu/Economics/bakija/BakijaTaxRatesAndLaborSupplyInOECDCountries.pdf
Since the tax *level* and economic growth literature is on the whole inconclusive, and the tax *structure* and economic growth literature generally finds that consumption taxes are growth enhancing, if labor taxes are theoretically equivalent to consumption taxes, then it stands to reason we should find little evidence that labor taxes have a negative impact on the quantity of labor supplied.
2. June 2014 at 08:14
Mark, A few comments.
1. The Alesina, et al paper starts right off saying Americans work much longer hours than Europeans. So if you are going to cite this paper, do you also agree with that claim? That was my claim, which you seemed skeptical about.
2. Micro labor supply elasticity estimates (discussed in Alesina) have no bearing on this issue, as they include an income effect. I agree that poor people tend to work longer hours.
3. I’m skeptical of unionization. France has extremely low unionization rates and very low hours. Admittedly it’s jus tone observation, but non-union hours also differ sharply.
4. Alesina points to “causes” like mandated holidays, but I see that as being really really hard to disentangle from taxes. They talk a lot about the importance of changes in social norms. OK, so how do we disentangle all this, when there are massively long and variable lags? How do we know that mandated holidays are just politicians catering to a public that wants to world less due to high taxes, and wants a fairly standard schedule to ease situations like school holiday/work holiday /vacation planning?
I’m not saying they are definitely wrong, just that a lot more work needs to be done.
I don’t agree that consumption taxes enhance growth, I agree that switching from capital taxes to consumption taxes enhances growth. But Europe added consumption taxes on to existing capital taxes (rates similar to US levels) to create higher overall taxes. That should reduce growth, and it seems it does. What other explanation do we have for low European GDP? Did their culture suddenly change radically after the 1960s, causing Europeans to prefer much longer vacations than Americans? I guess anything’s possible, but taxes are the theory to beat in my view. The one that matches theory with correlation.
I believe that at least 70% of taxes are labor taxes, properly defined. So I’m a bit confused when you suggest countries that collect 55% of taxes from labor have high labor taxes. What are the other 45%?
2. June 2014 at 14:56
Scott,
“So if you are going to cite this paper, do you also agree with that claim? That was my claim, which you seemed skeptical about.”
No, it’s the claim that taxes have anything to do with it, although Swedes now obviously work more than Americans, so some of what Alessina et al say in the opening paragraph is evidently no longer true.
“I’m skeptical of unionization. France has extremely low unionization rates and very low hours.”
The explanatory union variable that Alesina et al use is the union *coverage* rate. France has one of the highest union coverage rates in Europe.
“So I’m a bit confused when you suggest countries that collect 55% of taxes from labor have high labor taxes. What are the other 45%?”
In 2007 Switzerland collected 55.1% of its tax revenue from labor taxes, 22.4% from consumption taxes and 22.5% from capital taxes.
4. June 2014 at 12:13
Mark, I don’t see why any study would separate consumption taxes and labor taxes. Consumption taxes are also taxes on labor. I’d focus more on overall tax burden as a share of GDP.
8. January 2015 at 08:45
[…] via TheMoneyIllusion. […]