A simple model of division of labour [and comparative advantage] with unemployment | 175 comments (175 topical, editorial, 0 hidden)

Re: A simple model of division of labour [and comp (4.00 / 2)

As a member of a despised (here) caste of professional economists, I would say this is a 2-country, 1 factor of production partial equilibrium model. There are two sources of involuntary unemployment: the agents who are unemployable in both industries, and the ones who are not needed given the state of demand. The first type doesn't really matter for the model - just shift the origin in measuring the unemployment. The second type is there because you fix the demand exogenously.

Now, the basic models in international trade theory are at least 2-country, 2 factors. Yours is interesting because it has some distributions, but this doesn't really matter for the aggregate outcomes (other than the output isn't a linear function of employment, for example). All the logic will go through with equally productive agents. Of course, there are nontrivial policy implications if the government (or the 'social planner' as it's known in the literature nowadays) has the loss function over the number of unemployed, which is likely.

The model is partial equilibrium because it takes demand for both goods as given. With your timing assumptions (that your time tick is probably a quarter), even prices will have adjusted, and so the demands would have shifted. But I do understand that general equilibrium is ideologically impossible.

If you are really, really interested in what modern econo theory thinks about all of this, buy a used copy of Obstfeld-Krugman text. As they say, if you want to dispute neoclassical econ models, learn them first - what you've started already, BTW. Or there is a great on-line text on International Economics here.

by Sargon on Wed May 30th, 2007 at 06:32:00 PM EST

Re: A simple model of division of labour [and comp (none / 0)

Thanks.

The second type is there because you fix the demand exogenously.

I assume a given level of demand, but it should presumably be fixed endogenously. The model is incomplete, and I don't believe "agents ... who are not needed given the state of demand" would go away even if the demand were determined endogenously.

Now, the basic models in international trade theory are at least 2-country, 2 factors.

I thought Ricardo's basic model of comparative advantage was 2-country, 2-product, 1-factor? By the way, I'm thinking about a 2-factor model incorporating a Leontieff matrix to relate the products and the factors. Hopefully that will bring me closer to closure (endogenous demand).

Yours is interesting because it has some distributions, but this doesn't really matter for the aggregate outcomes [...] All the logic will go through with equally productive agents

You can't have a nonlinear PPF unless you have "some distributions". And a linear PPF is "not even wrong". Unequally productive agents really come into their own when you consider two countries and the possibility of "unemployment transfer", something I ran out of steam for when I was writing this diary.

Well, of course if there is no separation of time scales you cannot afford the kinds of "adiabatic approximations" implicit in my "basic assumption" and which I discussed in the subthread initiated by technopolitical. Is that what you mean by "general equilibrium is ideologically impossible"?

if you want to dispute neoclassical econ models, learn them first

I agree. I wonder whether I have the patience to read through a textbook any more (getting old, here) but Krugman comes highly recommended ;-) I see there's a new edition of the book just released in January of this year.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Thu May 31st, 2007 at 08:36:27 AM EST
[ Parent ]

Re: A simple model of division of labour [and comp (none / 0)

I don't believe "agents ... who are not needed given the state of demand" would go away even if the demand were determined endogenously.

As long as they are productive (probably requires wages are allowed to be a function of the type) and strictly prefer income from work to not working, they will be employed. You need "frictions" to get unemployment, and exogenously fixed demand is the standard one.

I thought Ricardo's basic model of comparative advantage was 2-country, 2-product, 1-factor?

I see. For me, "basic" means what I'd expect to teach in advanced MA or PhD class, and we don't do Smith or Ricardo these days. Economics is very different from philosophy, say.

You can't have a nonlinear PPF unless you have "some distributions".

Different technologies usually do the trick, but the only constant returns to scale technology with one input is linear, yes.

Is that what you mean by "general equilibrium is ideologically impossible"?

No, I mean the usual tone of statements at ET when people are talking about economics. Regretfully, a number of pretty sensible modeling assumptions and techniques were used in a heavily ideological way, and a number of people blame math instead of those having used it.

by Sargon on Thu May 31st, 2007 at 10:51:41 AM EST
[ Parent ]

Re: A simple model of division of labour [and comp (none / 0)

I thought Ricardo's basic model of comparative advantage was 2-country, 2-product, 1-factor?
I see. For me, "basic" means what I'd expect to teach in advanced MA or PhD class, and we don't do Smith or Ricardo these days. Economics is very different from philosophy, say.

See? The problem is that here we're debating what is published in the editorial pages of newspapers... And most often they just go on and on about how critics of Globalisation haven't understood Ricardo.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Thu May 31st, 2007 at 11:19:48 AM EST
[ Parent ]

Re: A simple model of division of labour [and comp (none / 0)

There you go assuming "full employment" as a given.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Thu May 31st, 2007 at 11:26:48 AM EST
[ Parent ]

Re: A simple model of division of labour [and comp (none / 0)

No. I'm just saying your model, as presented, doesn't allow for unemployment - if I extend it in a natural way. There are cozillions of ways to produce unemployment, of course, search models being the most popular lately. But you cannot do them in a simple semi-graphical model.

by Sargon on Thu May 31st, 2007 at 12:03:36 PM EST
[ Parent ]

Re: A simple model of division of labour [and comp (none / 0)

I'm just saying your model, as presented, doesn't allow for unemployment - if I extend it in a natural way

Can you spell that out?

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Thu May 31st, 2007 at 12:07:05 PM EST
[ Parent ]

Re: A simple model of division of labour [and comp (none / 0)

All the logic will go through with equally productive agents.

You cannot have division of labour with equally productive agents. Unless you assume that labour is homogeneous and capital is inhomogeneous, and that the origin of specialisation is the variation in the productivity of capital.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Thu May 31st, 2007 at 11:35:39 AM EST
[ Parent ]

Re: A simple model of division of labour [and comp (none / 1)

Or unless you have increasing returns to scale, in which case specialization might be path dependent based on historical accidents.

Are you talking about cross-country division of labor? In this case, standard story of different relative productivity (agents are homogeneous, with ratio of productivities in industry A/B different in the two countries) will generate nontrivial differences in labor allocation among industries.

by Sargon on Thu May 31st, 2007 at 12:09:28 PM EST
[ Parent ]

Re: A simple model of division of labour [and comp (none / 0)

Or unless you have increasing returns to scale, in which case specialization might be path dependent based on historical accidents.

In physics we call that spontaneous symmetry breaking. I thought there was more to job selection than spontaneous symmetry breaking.

Yes, that is assuming that different countries have homogeneous labour but for historical or geographical regions different capital or land productivities.

At some level it doesn't matter to me whether I'm talking about division of labour within one country or across countries. It's the same mechanism at work except that "country" is defined by mobility of capital and labour so that there is just one unit. But, then, if you have an international free market of capital and labour, is that not "like a country"? So, back to division of labour within a country.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Thu May 31st, 2007 at 12:23:22 PM EST
[ Parent ]

Re: A simple model of division of labour [and comp (4.00 / 2)

I still remember how we call this in physics :) Regretfully, economics is different - symmetries usually do not exist, and without them, how do you derive the conservation laws? Even Hamiltonians are discounted, which kills attempts to use a lot of established mashinery... and worst of all, the "molecules" are forward-looking - there the stat physics goes out of the window.

I really, really advice you to spend some time at Mark Thoma's Economist's View. Right now, there's a lively discussion of "heterodox" economics and current ways the neoclassical macro is changing. About a month ago, there was a series of posts on globalization's effect on labor market, and a very illuminating article by Krugman on modern international trade theory.

Finally, if neoclassics doesn't appeal to you (and general equilibrium, and especially rational expectations, do require some leaps of faith, I agree), why not looking into econophysics crowd? It would fit you very well, I guess.

by Sargon on Fri Jun 1st, 2007 at 05:17:55 AM EST
[ Parent ]

Re: A simple model of division of labour [and comp (none / 0)

I'll take a look at Mark Thoma's blog.

Now, you say "symmetries don't exist" and at the same time you were arguing above from the point of view of homogeneity of labour [that looks like a symmetry to me] and also said that you could just shift the origin [another symmetry] of the unemployment measure to exclude "unemployable" people.

Even if neoclassics doesn't appeal to me, I want to thoroughly comvince myself that neoclassical economics is bullshit before I write it off. I have some intellectual honesty left.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 1st, 2007 at 06:04:57 AM EST
[ Parent ]

Re: A simple model of division of labour [and comp (4.00 / 2)

Homogeneity of labor as a symmetry? Could be. And of course, we could simply shift the origin if we look at unemployed as equal (homogeneous) agents, exactly as they are considered in counting the unemployment rate. In this sense, they are symmetric. But as soon as you assume their productivities differ, at some of them (those who might work) aren't fully symmetric any longer.

The problem with this symmetry is - it's not "true". It's just a modeling assumption, which is relatively innocuous in some settings, but could backfire in others. Most of the time, it makes deriving results possible, and we just hope we aren't making an error which is too large. And, what's even worse, there isn't an obvious way of measuring the heterogeneity (and thus estimating the error) - how could I say that my "quality" is higher than someone else's?

It's hard for a physicist to accept that economics is a science. Economics is never going to achieve even 19th century physics' grasp of the "reality", I think. Still, I can see a progress - many people accept now that in 30 years, the profession could be very much different, and the ideological rigidity that has characterized economics for so long is slowly going away.

Well, best of luck to you from the other side of the pond - I guess I stopped being a physicist a long time ago...

by Sargon on Fri Jun 1st, 2007 at 08:48:50 AM EST
[ Parent ]

A simple model of division of labour [and comparative advantage] with unemployment | 175 comments (175 topical, 0 editorial, 0 hidden)

Top Diaries

Southern Irish disinterest in Northern Ireland affairs

The end is Nigh?

Leinster's demise and Munster's hope

Joe Biden. America's Interventionist Hero

The law of conservation of transness

Recent Diaries

Southern Irish disinterest in Northern Ireland affairs

Incident Ukrainian War Display Gets Int'l Attention

German Climate Activists Group Outlawed

Leinster's demise and Munster's hope

Joe Biden. America's Interventionist Hero

Targeted Assassinations: a Bankrupt Strategy

The end is Nigh?

Hijacking Peace by Zelensky for War Escalation

UA Drone Attack On Moscow

Rightwing Operation New EU 'Enlightenment’

Energy (and Other) Events Monthly - May 2023

May 2023

The law of conservation of transness

US Provoked Conflict In Europe

U.S. Will Be Ousted From Middle-East

Settler Extremism Blessed by Jerusalem Leaders

Occasional Series