Why Are Economists Frequently Surprised by Government Data?

[snoopies622]

Last night the CBS Evening News reported that the U.S. Government released some economic data about the month of November. One datum was that the net number of jobs in the United States increased by about 39,000. Their Senior Business Correspondent Anthony Mason said that that number "baffled economists, who were expecting a much stronger one." In the afternoon CNN also mentioned the newly released data, saying that "a lot of experts were shocked" by it.

I don't know how many times in my life I've seen or read news reports like this. We're told that the government released new economic data, and - usually in the same sentence - that economists were expecting something quite different and are therefore surprised.

These reports never describe or even name what economic models these experts used to make their (incorrect) predictions, why they had confidence in them in the first place, and which ones they'll be trying out next time.

I cannot help but wonder: is economics a science, or is it more like astrology? At least economists (unlike astrologers) are willing to admit when their predictions are false, but is the whole affair intellectually honest? One never hears, "last night starting around 8:31 pm there was an eclipse of the moon. Astronomers were shocked, since they predicted that it would come next Tuesday instead."

What's going on in those university economics departments? How can they be wrong so often? Where's the intellectual humility?

 

[talk2glenn]

You ever notice they never quote anybody? It's always "some economists" or just "economists".

Between you and me, nobody's really surprised by the results. However, what happens is the media will survey a number of university economists. To be counted, you have to give a number. Nobody has the means to accurately predict these numbers ahead of times. So yes, they're guessing.

But they are educated guesses. In this case, the consensus is that the unemployment picture in the United States is static. It isn't getting particularly worse (given population growth and seasonal fluctuations), and it isn't getting particularly better.

The "neutral" survey answer to "number of new jobs" is approximately replacement rate, or 130,000 new jobs. Expecting things to be slightly, but not meaningfully, better than this on account of the holidays, the surveyed economists generally revised their guesses upwards a bit. So you end up with a guess of 175,000, give or take, which should be read as"basically zero job growth".

What you got was 39,000. This is certainly worse than expected, but only marginally so. You're still basically at zero job growth, the uptick in published unemployment rates notwithstanding (again, it's probably seasonal - with the holidays coming you're apt to have people reentering the job market, but I haven't looked at the data so don't quote me).

It's sort of like the weather. A climate scientist can probably do a reasonably good job at telling you what direction the weather is going in (ie, will it be hotter or colder tomorrow, wetter or dryer) but a pretty poor job of telling you the particular temperature and/or humidity. Make sense?

 

[snoopies622]

...the media will survey a number of university economists. To be counted, you have to give a number.

So in other words, when CNN tells me that "economists were surprised.." they should instead say, "the average number (of new jobs, say) given to us by those economists who were willing to offer a prediction is different than what the government is now reporting." That would sound much better indeed.

I suppose what an economist would really like to offer is instead a range of possibilites, perhaps with a mean and a standard deviation, but that sort of option is not in the surveys.

 

[talk2glenn]

So in other words, when CNN tells me that "economists were surprised.." they should instead say, "the average number (of new jobs, say) given to us by those economists who were willing to offer a prediction is different than what the government is now reporting." That would sound much better indeed.

I suppose what an economist would really like to offer is instead a range of possibilites, perhaps with a mean and a standard deviation, but that sort of option is not in the surveys.

You've got it :)

The devils in the details. CNN isn't trying to mislead, it just doesn't want to make things more complicated than they need to be to get the basic idea across. And nobody ever wants to let the technical details spoil a good (and dramatic) story.

 

[snoopies622]

Still, I would find it much more interesting if we were told what formulae were used to make the predictions. Otherwise we (those of us who are ignorant of economics) are left wondering if it's all witchcraft.

 

[chingkui]

The "neutral" survey answer to "number of new jobs" is approximately replacement rate, or 130,000 new jobs.

Can you explain a bit what the "replacement rate" is and how do you estimate this number? Thanks.

 

[talk2glenn]

Can you explain a bit what the "replacement rate" is and how do you estimate this number? Thanks.

Sure; this is the number of new jobs that need to be created to offset the new entrants to the workforce due to population growth. If you have fewer new jobs created than this bumber, you have positive real job growth and negative net job growth.

It is a function of population growth and the labor force participation rate - the number of new Americans every month, and the proportion of those who enter the workforce. The former is basically constant (1%/year), while the latter is variable but predictable.

Currently, that rate is ~130,000 new jobs per month, give or take depending on methodology.

 

[snoopies622]

Nobody has the means to accurately predict these numbers ahead of times...It's sort of like the weather.

Do economists use computer models to make predictions like weather forecasters do? If so I would be happy to learn about them. Do particular models that are currently in use have names so I can go read about them?

 

[hamster143]

Economics is much more similar to climate science (and, unfortunately, even to philosophy) than to, say, physics. It works pretty well if you need to deal with supply and demand of widgets within the scope of a company X that sells them in city Y. Things are much more difficult to handle at the planetary level. There are many different, sometimes incompatible models all over the place. There's no clear cut mechanism for rejecting a bad model, because you can't really do clean experiments.

All sorts of factors enter at the planetary level, from random, self reinforcing movements in public sentiment (e.g. a random series of movements in the stock market resulting in a 5% drop over the course of a week will depress public morale, which will depress consumer spending, which will depress job creation ...), to weather, to actions of stock market / exchange rate / commodity speculators. It's hard to make a model that accounts for everything correctly. Especially when your model has emotional human beings as an intermediate step.

And the biggest problem is that economics is heavily influenced by politics. For example, if you were to ask ten major economists about some major economic event such as the early 2009 Obama stimulus and its impact on macro factors, you'd probably get answers ranging not only in magnitude but in sign. And those answers would correlate well with self described political affiliation of the economist. They tend to gravitate towards theories which are most convenient to them. An extreme example would be gold. We haven't been on a real gold standard in the developed world since 1933 or so, and most economists are inclined to think that gold standard is bad for economic health, but there's still a fringe (and maybe not even a fringe, depends on who you ask) that believes in it. It's like as if the minority of academic physicists believed in ether 80 years after Einstein. You could say that, after 80 years on fiat money, gold standard has not been conclusively "disproven" and it is unlikely to be.

Read this for starters, it should give you some insight

http://www.nytimes.com/2009/09/06/magazine/06Economic-t.html?pagewanted=all

In some way, I think, there is a distinction in economics between academic scientists and real-world scientists, similar to the one in computer science or physics. Academic scientists are paid to think up theoretical models, publish books, and teach students, regardless of having practical applications of their work. Real-world scientists are paid to have results. Academic scientists are polled by CNN to make predictions; real-world scientists work for Goldman Sachs, don't share their predictions with anyone, and use their predictions to trade on the stock market.

 

[snoopies622]

Thanks hamster143! That Times article looks very interesting.

And the biggest problem is that economics is heavily influenced by politics. ...answers would correlate well with self described political affiliation of the economist. They tend to gravitate towards theories which are most convenient to them.

That reminds me of this essay I read a few years ago about Milton Friedman, especially its opening paragraph:

http://www.wsws.org/articles/2006/nov2006/frie-n21.shtml

I suppose any field of study has to be funded by someone, and for specific purposes..

There's no clear cut mechanism for rejecting a bad model, because you can't really do clean experiments...It's hard to make a model that accounts for everything correctly.

What about isolating only two or three variables and looking for correlation?

 

[hamster143]

What about isolating only two or three variables and looking for correlation?

You can find correlations, but that's usually not enough to prove that some model is bad. Especially to people who are politically motivated to believe that their model is good and your model is bad. Someone will come out and say that there are factors X, Y, and Z, that you failed to consider, and that's why your explanation does not hold water. It's hard to gather the body of evidence to knock them out once and for all.

An interesting fact with regard to the statement I made earlier about Obama stimulus. Paul Krugman, author of the article I quoted (Nobel in Economics 2008), repeatedly argued that the stimulus was too small. Edward Prescott, referenced by the article (Nobel in Economics 2004), signed an open letter to Obama opposing the passage of said stimulus. Do you expect reliable consensus estimates from these people? It's best to follow a few of them and see whose predictions work out and whose don't.

 

[talk2glenn]

Do economists use computer models to make predictions like weather forecasters do? If so I would be happy to learn about them. Do particular models that are currently in use have names so I can go read about them?

Sure; there are a number of elementary models of varying complexity, including Heckscher–Ohlin, IS-LM, AD-AS, and Mundell-Fleming. These are just a couple; there are literally hundreds of individual models for any number of scenarios.

These models may then be incorporated together into a larger simulation, which is used for forecasting. Bernanke worked on an aggregate simulation called the "financial accelerator", for example, and most major banks and brokerages have their own simulations. Unfortunately, they really aren't all that accurate, speaking specifically. They are relatively useful for making general claims, but not particularly useful for telling you the date and time at which particular events might occur.

The accuracy of economic forecasting increases as scale decreases. For example, with enough data, we can predict almost perfectly the effect of a change in the price of, say, electricity on a population of consumers, the utility, and community social welfare. We can predict the effect of the institution of new taxes (revenue raised and social welfare lost). We can advise startups on what price to charge and how much to produce when entering a new market.

This is all quite possible, using relatively simple functions. You might try looking at an introductory econometrics textbook; that should have an appendix with hundreds of models to entertain you.

Economics is much more similar to climate science (and, unfortunately, even to philosophy) than to, say, physics. It works pretty well if you need to deal with supply and demand of widgets within the scope of a company X that sells them in city Y. Things are much more difficult to handle at the planetary level. There are many different, sometimes incompatible models all over the place. There's no clear cut mechanism for rejecting a bad model, because you can't really do clean experiments.

Is this so different from physics? In principle, a physicist can (given sufficient data) accurately predict the behavior of physical systems ad infinitum. In practice, things don't quite work out that way. There are externalities and variables whose prediction is either impossible, given the state of physics, or faulty. Physics cannot perfectly explain the universe, and even those things which can be explained are still subject to input error. Economics is no different; it cannot perfectly explain markets, and even those things which can be explained are subject to input error. However, it is quite good at explaining lots of things.

And the biggest problem is that economics is heavily influenced by politics.

I disagree. Economics is, fundamentally, not political. An economic claim is either accurate or inaccurate. It's application is, fundamentally, political. Economics can tell you that there are scarce resources to meet unlimited demands. It cannot tell you how to distribute those resources, because there is no right way and no wrong the way. The former is an economic problem, the latter a political one.

An extreme example would be gold. We haven't been on a real gold standard in the developed world since 1933 or so, and most economists are inclined to think that gold standard is bad for economic health, but there's still a fringe (and maybe not even a fringe, depends on who you ask) that believes in it.

There is no disagreement amongst mainstream economists about the consequences of a gold standard. You seem to be approaching this subject from the position that there is a "right" and "wrong" policy answer to an economic problem. This is false, just as it is false in physics. A physicist can teach you the principles of rocketry; the application of that principle is a political answer. From the perspective of physical science, there is no right or wrong way to use rocketry - bombs or spaceships. This is policy versus theory.

It is economic fact that a basis of trade is a more efficient mechanism for exchanging resources in a market than the barter system, and this can be demonstrated experimentally. Different trade bases have inherent advantages and disadvantages, but there is no right answer. How currency is introduced to a market is a policy decision. Going with the gold standard (or any standard which relies on fixed exchange rates) has the advantage of a stable asset value, and the disadvantage of disallowing an independent monetary policy and a higher risk of deflation (you're trading inflation risk for deflation risk basically - neither is desirable). Politicians must make an informed decision about what is best for their economy.

In practice, a free floating currency is generally preferred in developed, stable economies that aren't at significant currency risk, because is frees that economy to implement an independent monetary policy. On the other hand, fixed exchange rates are generally preferred in developing economies, because it eliminates some risk of destabilization inherent to weaker economies, and increases the countries competitiveness in the international trade and investment markets.

 

[hamster143]

You seem to be approaching this subject from the position that there is a "right" and "wrong" policy answer to an economic problem.

It's more than that. There's the bigger question of the business cycle and the effect on the business cycle of things like gold / fiat money, fiscal policy, and monetary policy. Did the Fed cause the Great Depression (Friedman), or was it not active enough to prevent the Great Depression (Krugman)? If that is not settled, how can be expect to have a reliable consensus about down and dirty things like the number of new job openings in a week?

Back on the subject of gold standard, one well known economist whose views on the relationship between the gold standard and the business cycle were quite distinct from both Friedman and Krugman was Friedrich Hayek (Nobel in Economics 1974), who happened to be a professor at the University of Chicago (among other things), and, even twenty years in the grave, he has the ear of at least one newly elected congressman. So, does ether ever die in economics?

 

[talk2glenn]

It's more than that. There's the bigger question of the business cycle and the effect on the business cycle of things like gold / fiat money, fiscal policy, and monetary policy.

What question? This has been answered definitively.

Did the Fed cause the Great Depression (Friedman), or was it not active enough to prevent the Great Depression (Krugman)?

Both Krugman and Friedman agree, as does the principle, that the actions of the Fed during the Depression increased the effective reduction in equilibrium output.

I have attached a simple MS Paint sketch by yours truly illustrating the point. Whether the Fed was right or wrong to try and maintain a target interest rate, regardless of the consequences in terms of output, is again a policy question and not an economic question.

ack on the subject of gold standard, one well known economist whose views on the relationship between the gold standard and the business cycle were quite distinct from both Friedman and Krugman was Friedrich Hayek (Nobel in Economics 1974), who happened to be a professor at the University of Chicago (among other things), and, even twenty years in the grave, he has the ear of at least one newly elected congressman. So, does ether ever die in economics?

Hayek's assertion that an increase in the money supply reduces effective interest rates is neither controversial 'nor a topic of debate, but point of fact. Again, the interpretive consequences of this effect is not a factual question. Once you accept that a liquid money supply poses inflationary risks, you must make an informed decision regarding an economies currency vehicle. Hayek has an opinion, but he cannot definitively state which approach is practically superior. There are risks, costs, and benefits associated with any policy decision.

We do the best we can to make the decisions which present the least risk and offer the greatest rewards, given the data.

 

[hamster143]

What question? This has been answered definitively.

Evidently not, otherwise the 2004 and the 2008 Nobel laureates would not be expecting two opposite outcomes of fiscal stimulus in the present economy. The problem seems to be that the 2004 laureate firmly believes in the Ricardian equivalence (and the truckload of mathematical theories that rest on it), and the 2008 laureate considers it only approximate even in the best of times. Which is a pretty significant and broad-reaching theory, and hardly a new one, so you would've expected there to be a consensus.

Also on the subject of the present economy, a curious read:

http://siteresources.worldbank.org/INTMACRO/Resources/CaseyMulligan_041609.pdf

again a professor of economics at UofC. Here's the gist of it. Labor input is way down since the beginning of the recession, and per capita productivity is way up, and the increase in productivity is greater than the author would've expected, given the changes in labor. The author does some mathematical trickery with his models (I can't say I follow this part very well, he seems to suggest that, in his models, a reduction in consumption, ceteris paribus, is supposed to lead to an increase in labor) and concludes that there's a mysterious external "labor supply distortion" in play, which is responsible for more than 100% of the employment decline since December 2007. He then hypothesizes that people choose to stay out of work because "When all of the instances of means-tests [e.g. the ability to refinance your house if you're struggling to make your payments. -h.] are considered in combination, a number of workers in the U.S. economy may have a terrible incentive to work." And that's why we have 10% unemployment in the country.

And before anyone goes and tries to find some harsh words for this scientist and this gem of modern economic thought, keep in mind that the article was, apparently, reviewed by Gary Becker (Nobel in Economics 1992, a yet another professor at UofC) and found worthy of publication.Now, there's obviously a vast chasm between this approach and the neo-Keynesian views of Krugman & co. And if you go to the guy's blog, the first entry as of today has the words "Big Government" in the title, so you could pretty much guess right away how he voted in the last election.

 

[talk2glenn]

Evidently not, otherwise the 2004 and the 2008 Nobel laureates would not be expecting two opposite outcomes of fiscal stimulus in the present economy.

Repeating your claim ad infinitum while changing the names of the principles does not change the facts. The apparent dissension is an analytical error on your part; you've simply misunderstood the nature of the debate. This is not unreasonable - if I picked up a technical paper on theoretical physics, I might conclude based on the contention and argument that there is no consensus on anything amongst physicists. In actuality, all parties to the debate agree on fundamental principles, and debate on what follows there from (theory versus law).

Given your failure to regard my points, I won't address this claim further.

Here's the gist of it. Labor input is way down since the beginning of the recession, and per capita productivity is way up, and the increase in productivity is greater than the author would've expected, given the changes in labor.

That is not the gist of the authors claims. Your analysis is quite flawed - you've chosen to analyze a technical paper without a technical basis, and not surprisingly, you've failed.

The labor market can be defined as the pursuit of equilibrium in the labor supply and labor demand curves, with time (input hours) being a function of the wage rates. The labor supply curve itself can be thought of as the series of equilbria points in the consumer demand for labor relative to leisure. At equilibrium, the ratio of the prices (P(w)/P(l)) is equal to the marginal rate of substitution for the consumer between work and lesiure. That is, for a given wage level, the consumer is willing to work a certain number of hours and play a certain number of hours such that the last hour spent working gives him as much utility as the last hour spent playing. This is true for every point on the labor supply curve; each point represents a case where MRS = P1/P2. Likewise, for the labor demand curve, each point can be thought of an equilibrium level where the marginal product of labor is equal to the price rate (the value of the last unit of labor is equal to the value of the last unit of production).

When the two curves are plotted relative to one another on the same graph, an equilibrium level of output is found. I have sketched this for you and attached it here, with that equilibrium being defined as Y*. By definition, because each of those points represents a given value for MPL and MRS, respectively, then it must be case that where they intersect, MPL = MRS. This is labor market equilibrium.

During this recession, the MPL has risen, clearly (this is a function of employer response to recession - they eliminate their most expensive, and least profitable employees). However, the author argues that the MRS has, conversely, fallen. Look at the formula he has given to approximate MRS. MRS is consumer willingness to substitute labor for leisure and vise versa while remaining just as happy. When labor is behaving normally, a decline in wage rates makes leisure cheaper relative to labor - consumers respond by buying less labor and more leisure, but consumers are also poorer, and they respond to a decline in income by consuming less leisure and more labor. Author argues that the substitution effects dominates the income effect, and MRS is declining (it is taking less leisure time to compensate workers for the loss of an hour of labor time). This is occurring even as MPL is rising (and employer willingness to pay for that labor is also rising).

This has produced a disequilibrium in the labor markets (which I have illustrated as the vertical line in the graph). This condition is not sustainable, and eventually one of three things must happen. The macroeconomic market can recover to pre-recession conditions, and the labor market will return to the Y* equilibrium level. Consumers can respond by permanently reducing the number of hours they are willing to supply to the labor market at any given wage level (moving the labor supply curve to the left). Producers can respond by permanently reducing the number of labor hours they are willing to hire at any given wage level, moving the labor supply curve to the left.

The author argue that consumers, at least for now, appear to be responding by shifting their labor supply curves inward. When the author uses the word consumption, keep in mind the equivalence between consumption and income. If he says that consumption is falling, he is also saying that income is falling. Consumers are effectively reducing the supply of labor they are providing to the market, and consuming additional leisure. The labor supply curve is shifting left. As the curve moves, an opportunity for a new equilibrium output level below Y* but above the current, recessionary output level is possible. The market should be trending towards these equilibrium levels at each opportunity (hence is claim that a reduction in consumption would be followed by an increase in the supply of labor - this is a correlative fact, not a causal fact, as you thought). That it is not implies that the consumer labor market will continue to contract, until an equilibrium is achieved.

The author then tries to explain why consumers are behaving in this manner (what is prompting them to reduce their value of labor and increase their value of leisure). He posits that welfare and government assistance benefits contribute, but that a driving factor is mortgage modification. Bank willingness to modify consumer loans is a function of consumer ability to pay - the more able to pay, the less willing to modify. Employment is associated with increased ability to pay.

I apologize for the length, and hope this was clear, but my point here is to demonstrate that nothing in the authors position is controversial or in contradiction with economic principle. The author is using those principles to try and explain behavior in the labor markets. I do not necessarily agree with the authors position, but at face value the road to that conclusion appears sound.

The author does some mathematical trickery with his models (I can't say I follow this part very well, he seems to suggest that, in his models, a reduction in consumption, ceteris paribus, is supposed to lead to an increase in labor) and concludes that there's a mysterious external "labor supply distortion" in play, which is responsible for more than 100% of the employment decline since December 2007.

The "labor supply distortion" is not a mysterious force at all. The author himself explains that this is simply his chosen identifier for the difference between MRS and MPL and the delta in that difference observed during the recession (up to the time of writing). This is a disequilibrium that is being addressed (in the authors opinion) by a leftward movement of the labor supply curve.

Just because you do not understand the mathematics, does not mean that it is trickery.

Now, there's obviously a vast chasm between this approach and the neo-Keynesian views of Krugman & co. And if you go to the guy's blog, the first entry as of today has the words "Big Government" in the title, so you could pretty much guess right away how he voted in the last election.

No, there isn't. Both parties agree on the underlying principles. The parties may disagree on which effects and forces are dominating, but they cannot both be correct. Ultimately, econometric analyses will determine what held, where, and why.

An interventionist like Krugman might argue that, given this picture, the government should intervene to stimulate production demand, such that the labor demand curve is shifted temporarily to the right until a satisfactory equilibrium point above the current level is reached. A non-interventionist might argue that assistance and support programs for consumers are eliminated, such that the relative value of labor to leisure increases (it becomes more expensive to give up an hour of labor for an hour of leisure). Both are means to an end, and we come full circle - policy versus principle.

 

[hamster143]

In actuality, all parties to the debate agree on fundamental principles

Do all parties agree on Ricardian equivalence?

That is not the gist of the authors claims. Your analysis is quite flawed - you've chosen to analyze a technical paper without a technical basis, and not surprisingly, you've failed.

I admit, I do lack a technical basis in neoclassical thought, and I was too sleepy yesterday. The MRS plays a larger role than I thought.

Consumers can respond by permanently reducing the number of hours they are willing to supply to the labor market at any given wage level (moving the labor supply curve to the left). Producers can respond by permanently reducing the number of labor hours they are willing to hire at any given wage level, moving the labor supply curve to the left.

One small point: consumers respond by reducing labor supply, and producers respond by reducing labor demand. While in the equilibrium the two should be equal, downward shifts of labor supply curve and labor demand curve both have the effect of contracting labor market. In fact, the big problem of the article is that the author concludes that there's an external effect on labor supply (because MRS of labor/leisure is down?), when it is really labor demand that should be implicated.

Look at the formula he has given to approximate MRS. MRS is consumer willingness to substitute labor for leisure and vise versa while remaining just as happy.

This seems to be the starting point of the chain of misconceptions in the article. I have to go now so I'll write about it later.

but my point here is to demonstrate that nothing in the authors position is controversial or in contradiction with economic principle.

I did not deny that. There is a subset of economic principles that leads us to conclude that 5% of Americans choose to remain unemployed and live on $400/week UI paychecks in hopes of having the government help refinancing their mortgages. There's a different subset that leads us to conclude that there's a shortage of aggregate demand, and that's why those 5% can't find jobs, even though they try hard. They are all sound economic principles, and, for the most part, none of them are in contradiction with each other.

No, there isn't. Both parties agree on the underlying principles. The parties may disagree on which effects and forces are dominating, but they cannot both be correct. Ultimately, econometric analyses will determine what held, where, and why.

If we had universal and commonly agreed underlying principles, these principles would tell us which effects are dominating, we wouldn't have to wait for the current generation of economists to die to get the answers. Instead, we have models used by Krugman and totally different models used by Mulligan, and these models come to different conclusions about the effects. Which was my point from the beginning. Furthermore, my problem is that we get apparently respected economists publishing papers that make utterly ridiculous claims which don't even pass the reality check - like this one - which puts this whole class of models into question.

 

[mheslep]

I cannot help but wonder: is economics a science, or is it more like astrology? At least economists (unlike astrologers) are willing to admit when their predictions are false, but is the whole affair intellectually honest? One never hears, "last night starting around 8:31 pm there was an eclipse of the moon. Astronomers were shocked, since they predicted that it would come next Tuesday instead."

It seams to me a better analogy would be to weather prediction, about two weeks out. Such predictions are often mistaken, but this doesn't mean the science underlying the predictions is all crack pottery.

 

[hamster143]

Ok, now that I'm fully awake, I think I can explain what went wrong in that article.

He approximates MRS of labor/leisure as equal to the ratio of real consumption to leisure time. This by itself is somewhat vague and oversimplified (how many leisure hours does an average full-time worker have? How do we count leisure hours of people out of labor force? Shouldn't people employed part-time have a higher preference to labor than people employed full-time?) but manageable. But the real problem in my view is that he does not add savings to consumption. Not being a professor of economics myself, I can't offer the textbook justification for that, which may exist, but it clearly shouldn't apply in this case. Lower consumption to leisure ratio makes it appear that people are less likely to work ("labor supply distortion"), when in reality they are more likely to save.

If I crunch the numbers, guessing the value of T that the guy uses, I do, in fact, get that MRS excluding savings is down 3.5% from 1/2008 to 5/2009. But, if I include savings, MRS goes up 2% over the same period of time, almost as much as productivity. So that turns all the findings upside down.

Of course, if we want to bring savings into the picture, that's a whole new can of worms, since we now have to think about money supply and third-party forces, because the situation where everyone hoards 5% of their income as cash permanently is not an equilibrium state. Is that employment model even equipped to handle sharp changes in the saving rate? ... It's possible that, when the dust settles and the curves are done moving around, we will in fact end up with the gap between MPL and MRS, which is not caused by a labor supply distortion beginning in Q4/2007, but by a labor demand distortion beginning in Q2/2009.

Furthermore, from what I was able to glean from the textbooks, the whole idea that the personal saving rate could have risen in 2008 is incompatible with neoclassical economics 101, because higher saving is supposed to occur when households engage in intertemporal substitution of consumption, and that happens when real interest rates (or their expectations) go up. To quote, "the marginal rate of substitution between consumption this period and next must equal ... the marginal rate of return." But in 2008 real interest rates were down. Beyond economics 101, it seems that there are multiple different models of the saving rate. At least one author admits, "the time series evidence for intertemporal substitution in consumption is far from definitive." In practice it probably means that some people will use their favorite models, and most will assume that the saving rate is either constant or proportional to the short-term real interest rate, and, when the saving rate makes an unexpected jump, many analysts will be surprised.

 

[snoopies622]

It seems to me a better analogy would be to weather prediction, about two weeks out. Such predictions are often mistaken, but this doesn't mean the science underlying the predictions is all crack pottery.

Indeed. But then one also never hears, "Today in Washington it rained. Meterologists were shocked."

 

[skilgannonau]

Indeed. But then one also never hears, "Today in Washington it rained. Meterologists were shocked."

I think this is theatre on the part of the news networks. I'm sure there were economists that weren't shocked. It's probably already been discussed but economics is not physics so I don't think the same standards, e.g. prediction, should be applied. As far as we know there is no fundamental force exist that govern social interaction as there exists in the physical world that govern physical interactions. Hence, economists have been reduced to identifying relationship between variables that approximates what we actually see in the real world. This is hard work which isn't always fruitful. Hence why in modern macroeconomics the paradigm has changed half a dozen times in the last 70 years. I've stated this in another post, but unless you a see fundamental breakthrough in economics, economics isn't going to have the same predictive power as there is in physics (or even weather science). At best you're going to have approximations.

 

[hamster143]

Cochrane’s heaviest punching is reserved for Krugman’s support for President Barack Obama’s fiscal stimulus. He invokes the hoary old “Ricardian equivalence theorem,” revived by the Harvard economist Robert Barro, according to whom “debt-financed spending can’t have any effect, because people, seeing the higher future taxes that must pay off the debt, will simply save more. They will buy the new government debt and leave all spending decisions unaltered.”

In short, Krugman “has absolutely no idea about what caused the crash, what policies might have prevented it, and what policies we should adopt going forward” – except that the government should now spend money like a drunken sailor. Far from having too much math, economists need more, in order to “keep the logic straight.”

On the stimulus, though, Krugman achieves a knock-out punch. The view that extra government spending “crowds out” an equal amount of private spending, so that its net stimulus effect is zero, would be true only if the economy were at full employment. Indeed, the Chicago School tacitly assumes that economies are always at full employment. They are unfazed by the fact that America’s economy has shrunk by 4% in the last year and that over 6 million people have been added to the unemployment rolls.

To Chicago economists, an increase in the number of idle workers represents a voluntary choice for leisure. In a concession to commonsense, they concede that people may make mistakes and, to that extent, a stimulus may be beneficial. But they insist that the only stimulus that will work is printing money. This will drive down interest rates and lead to an economic rebound.

Against this view, Keynes pointed out that interest-rate reductions may not do the job, because at zero or near-zero rates investors hoard cash rather than lend it out.

http://www.project-syndicate.org/commentary/skidelsky22/English

"Both parties agree on underlying principles" my foot ...

By the way, every person who reads this who voluntarily reduced his/her spending in anticipation of higher future taxes in order to pay for Obama stimulus, raise your hand.

 

[talk2glenn]

He approximates MRS of labor/leisure as equal to the ratio of real consumption to leisure time.

He calls it approximately equal to the ratio of consumption per person to leisure hours. Again, this is not a technical definition of the MRS between labor and leisure, but it will suffice as an approximation. Consumption is roughly equivalent to income in an economy (once you account for lending and investment), when measured economically (don't confuse economic savings with savings in finance) - it's not perfect, but the margin of error will be pretty small.

This by itself is somewhat vague and oversimplified (how many leisure hours does an average full-time worker have? How do we count leisure hours of people out of labor force? Shouldn't people employed part-time have a higher preference to labor than people employed full-time?) but manageable.

1 - (ratio of time worked to total time) = proportion of time spent not working, or leisure time. Whether one is working or not is irrelevant. Someone not working is understood to have a MRS that precludes work - the cost of working is too high relative to the value of leisure (and considering that value of that persons labor, of course). We don't differentiate between the unemployed and the employed here. Every living individual has a propensity to work, and a propensity to leisure. Whether they do or not not in practice is a function of their individual values. To be more specific, you do not differentiate between people "inside" and "outside" the work force when defining aggregate MRS between labor and leisure.

Is this clarifying or confusing?

Of course, if we want to bring savings into the picture, that's a whole new can of worms, since we now have to think about money supply and third-party forces, because the situation where everyone hoards 5% of their income as cash permanently is not an equilibrium state.

Generally speaking, this kind of hoarding is highly improbable. If you imagine a state of disequilibrium between savings and investment (savings falls below investment), then interest rates will rise - the demand for investment capital exceeds supply, forcing bond issuers to lower prices, with interest rates moving in the opposite direction of price. This provides a powerful disincentive to hoard.

In practice, people tend to save less, not more during recessions, because cash is a normal good - generally speaking, the richer you are, the greater your demand for cash, and the poorer you are, the smaller your demand. This seems counterintuitive at first, but in practice is quite logical.

Think of it in aggregate, and not individually. Cash is used to buy things. Hold the money supply constant, but imagine that the people holding that money are now poorer (the economy is in recession). This reduces their demand for things, reducing the demand for cash. The effect is to increase the demand for bonds, and interest rates fall.

Furthermore, from what I was able to glean from the textbooks, the whole idea that the personal saving rate could have risen in 2008 is incompatible with neoclassical economics 101,

Again, you must differentiate between "savings" in finance and saving in economics. Savings in finance refers to the low-risk preservation of money. Savings in economics refers to a change in the demand for money, either on the investment or consumption side. Assuming no change in the money supply, a decrease in one leads to an increase in the other.

One small point: consumers respond by reducing labor supply, and producers respond by reducing labor demand.

You are correct, of course. That was a typo.

 

[hamster143]

Generally speaking, this kind of hoarding is highly improbable. If you imagine a state of disequilibrium between savings and investment (savings falls below investment), then interest rates will rise - the demand for investment capital exceeds supply, forcing bond issuers to lower prices, with interest rates moving in the opposite direction of price. This provides a powerful disincentive to hoard.

In practice, people tend to save less, not more during recessions, because cash is a normal good - generally speaking, the richer you are, the greater your demand for cash, and the poorer you are, the smaller your demand. This seems counterintuitive at first, but in practice is quite logical.

that is true, but that is exactly what happened in 2008:
http://research.stlouisfed.org/fred2/graph/?chart_type=line&s[/url][1][id]=PSAVERT&s[1][range]=5yrs

 

[talk2glenn]

that is true, but that is exactly what happened in 2008:
http://research.stlouisfed.org/fred2/graph/?chart_type=line&s[/url][1][id]=PSAVERT&s[1][range]=5yrs

Again, you need to differentiate between savings as used in finance and as used in economics. Forget everything you "learned" reading news stories about the BLS personal savings rate, and consider what I told you.

The PSR is loosely defined as income - taxes - debt repayment - consumption.

However, the question you need to be asking yourself is what happens to that balance? It is not hoarded, as implied by the news stories on the savings rate; it is leant out. Somebody else borrows that money, and uses is to consume. So one mans "savings" is another mans "consumption", and in aggregate, income = consumption. I'm throwing terms around loosely. More technically, income = consumption + investments, but you get the idea.

To understand what is happening economically to the demand for money (economic savings rate), you need to look at the change in interest rates, and correct for changes in fiscal policy. Corrected interest rates fell during the recession; the economic savings rate fell too. You can also look at bond prices; they too rose during the recession, confirming the hypothesis.

Look at the 3rd component of the PSR. Debt repayment. Where'd that debt come from, and what did it finance before you paid it back?

EDIT: You could also take another look at the graph I threw up on page 1, relating to the depression. When output fell during the initial economic contraction, the response was a decrease in interest rates caused by a reduced demand for savings (cash), prompting the Fed to tighten fiscal policy to maintain the target rate. This is a good graphical illustration of the concept. This is also why classical economists reject the need for government intervention to correct recessions; the market is essentially self-correcting between the personal consumption markets and the money markets. A drop in the one (recession) causes an expansion in the other, fueling future investment growth in the former (with individual self correcting mechanisms built in; falling consumer prices entices people to start spending money again, and falling interest rates encourages them to stop buying bonds on the other side).

Post-Keynesian theory complicates things a bit, but the concept is sound.

 

[snoopies622]

...economics is not physics so I don't think the same standards, e.g. prediction, should be applied. As far as we know there is no fundamental force exist that govern social interaction as there exists in the physical world that govern physical interactions.

This reminds me of "psychohistory" from the Asimov Foundation series. The premise goes something like, although one cannot predict the behavior of an individual, one can predict the behavior of a large group, and the larger the group the more accurate the prediction.

Fiction, of course...

 

[talk2glenn]

I've stated this in another post, but unless you a see fundamental breakthrough in economics, economics isn't going to have the same predictive power as there is in physics (or even weather science).

Define your expectations. If the price of movie tickets rises by 1% tomorrow, consumers purchases will fall by approximately 0.9%. This is confident predictive power. Economics has the disadvantage of transformity. The price elasticity of demand for movie tickets may vary over time, in response both to variables of economic interest and also to variables of no direct economic interest (changes in consumer tastes; the availability of substitutes). But the underlying assumption - that consumers will demand film in quantifiable rates relative to its price - will not.

Economics is sufficient to explain resource-maximizing behavior conceptually, and predict it practically. The former is unconstrained and theoretical; the latter dependent on the availability of data and the consistency of preferences. But this is no different in practice than any other science. A physicist holds constant the mass of the Earth when calculating its gravity. From that derivation, a great deal more follows. The only distinction is the rate of transformity - a physicist can count on the mass of Earth changing only fractionally moment to moment. An economist lacks that luxury.

Now, you might argue that a physicist can say, with a certainty, what effect a 25% overnight mass loss event might have on Earth's gravitational pull, and from there tell me lots of stuff about architectural engineering, gravity, etcetera. So could an economist. But you fault the economist because he cannot tell you when or even if the (metaphorical) mass of the Earth will change by 25%.

News flash - neither can the physicist. Economists are held to an impossible standard. Predictive science is based on repeatability of results when inputs are held constant. Economics can tell you a great many things, and make a number of useful and repeatable predictions. It does not need to be immutable (the fact that it's changed 7 times in 70 years has no relationship), and neither does the tested environment.

 

[skilgannonau]

This reminds me of "psychohistory" from the Asimov Foundation series. The premise goes something like, although one cannot predict the behavior of an individual, one can predict the behavior of a large group, and the larger the group the more accurate the prediction.

Fiction, of course...

But that was never my premise ...

 

[snoopies622]

But that was never my premise ...

Oh, I know. I just said that what you said reminded me of it.

 

[skilgannonau]

If the price of movie tickets rises by 1% tomorrow, consumers purchases will fall by approximately 0.9%. This is confident predictive power. Economics has the disadvantage of transformity. The price elasticity of demand for movie tickets may vary over time, in response both to variables of economic interest and also to variables of no direct economic interest (changes in consumer tastes; the availability of substitutes). But the underlying assumption - that consumers will demand film in quantifiable rates relative to its price - will not.

Agree.

‘[T]he underlying assumption - that consumers will demand film in quantifiable rates relative to its price’ is observable and we’ve used the fiction of individual utility to provide a theoretical foundation for this observation. This has been my point this whole time. A utility function can depend on any infinite number of variables. This means ‘economists have been reduced to identifying relationship between variables that approximates what we actually see in the real world.’ Trying to identify which of the infinitely many variables (and the type of relationship between variables) is relevant is ‘hard work which isn't always fruitful’. The social world isn’t like the natural world where you have a simple relationship like for instance F = ma.

Economics is sufficient to explain resource-maximizing behavior conceptually, and predict it practically. The former is unconstrained and theoretical; the latter dependent on the availability of data and the consistency of preferences. But this is no different in practice than any other science. A physicist holds constant the mass of the Earth when calculating its gravity. From that derivation, a great deal more follows. The only distinction is the rate of transformity - a physicist can count on the mass of Earth changing only fractionally moment to moment. An economist lacks that luxury.

Agree. See my point above about identifying the relevant variables and the relationship between those variables.

Now, you might argue that a physicist can say, with a certainty, what effect a 25% overnight mass loss event might have on Earth's gravitational pull, and from there tell me lots of stuff about architectural engineering, gravity, etcetera. So could an economist. But you fault the economist because he cannot tell you when or even if the (metaphorical) mass of the Earth will change by 25%.

I’m not faulting the economist. This is what I don’t get. You are making the same point I’m making but in different words:

News flash - neither can the physicist. Economists are held to an impossible standard.

I’m saying the same thing as you are: You can’t set the same standards for economists and physicists with respect to prediction. Physicists have an easier job because of the fundamental forces.

Predictive science is based on repeatability of results when inputs are held constant. Economics can tell you a great many things, and make a number of useful and repeatable predictions.

Agree.

It does not need to be immutable (the fact that it's changed 7 times in 70 years has no relationship), and neither does the tested environment.

Agree that it doesn’t need to be immutable but I only mentioned the fact ‘it's changed 7 times in 70 years’ to illustrate the point the ‘science’ of macroeconomics is hardly settled (because the relationship between in the social world a lot more difficult to discern than it is in the natural world). It was never an argument that economics should be ‘immutable’.

Look I’ve only skimmed read most your posts (so advanced apologies if I misrepresent you) but I’m getting the impression that you probably have a lot more confidence in the state of macroeconomic theory than I probably have. Fair enough. But it seems to me we are saying the same things with respect to economic prediction (even if it is in a different manner).

 

[skilgannonau]

Oh, I know. I just said that what you said reminded me of it.

Ok, I get it now. You meant 'fiction' literally as opposed to metaphorically. Cheers.

 

[WhoWee]

Can you think of a field of study that has more variables, so few constants, and so many manipulative participants?

 

[snoopies622]

No. And yet in principle one should be able to apply the scientific method to it, by isolating variables and looking for correlations. Maybe it just needs another few hundred years.

 

[WhoWee]

No. And yet in principle one should be able to apply the scientific method to it, by isolating variables and looking for correlations. Maybe it just needs another few hundred years.

The scientific method should be applied to health care reform.

 

[bobc2]

...I've stated this in another post, but unless you a see fundamental breakthrough in economics, economics isn't going to have the same predictive power as there is in physics (or even weather science). At best you're going to have approximations.

Wait... Weather science IS physics. The system is quite complex and nonlinear but nevertheless modeled by physical components which obey laws of physics.

Physicists developed finite element models (sometimes 1,000,000 x 1,000,000 matrices) of some of the Space Shuttle Orbiter subsystems to predict dynamic stresses and strains across the structures during ascent (mechanical vibrations). They at least had the advantage of testing the structures on the ground in order to adjust the models to produce the same vibrational resonance frequencies and other modal parameters. And even after all of that pretesting, the predictions of stresses and strains were in error by around 10% for certain modes of vibration, and in error of over 100% at high frequencies.

Just to make the point that even though F = ma is well established, there is no guarantee that the model for a given physical system will be simple. By the way, physicists don't even try to apply closed form solutions to predict interactions between the very complex molecules.