Asymmetrical Information

Wait a minute. So Jane Silk is interviewing Jane Galt? I think I'm going to change my name to Jane Johnson.

In 1992, total state and local spending has 10% of GDP. It steadily fell from there down to a nadir of 9.4% of GDP in 1998, stayed there for a year, and then steadily climbed back up to 10.1% in 2002.

What spending binge?

Why should spending be a function of GDP, rather than population+inflation?

Jane asked:
"Why should spending be a function of GDP, rather than population+inflation?"

'Cos increases State workers compensation will track with increase with private-sector personal incomes (which will track with productivity increases), if States are to remain competitive in the labour market.

Number of state employees would be expected to track with population increase (if levels of service are kept constant). Productivity growth + population growth = GDP growth.

So, did you do any quantitative research on state & local spending before appearing on the programme, or did you just wing it?

So why is there an inverse correlation between the number of state workers in NYS and the state's population?
And state employees have benefits that don't factor well into comparing public vs. private sector, such as retirement, a large number of vacation days, and "tenure." It is very difficult to document and fire a state worker, and a lot of the workers know that. I'm not saying that all state workers fall into this category, but a fair number of them do. I had some excellent workers when I was chief clerk for a city court in NYS, but I also had some slackers, and there was not much I could do about the slackers. In the private sector, I could fire them and hire other workers, but in the public sector, that is not really an option.

"Why should spending be a function of GDP, rather than population+inflation?"

To pick a random state, here's Nebraska. 22% goes to education. I think that's mostly salary, which should track with GDP. 31% goes to health & human services. I think that's code for Medicaid, which I'd expect to track GDP growth. General government is probably mostly salaries, 9%. That's 62% right there.

The rest is in roads and the justice system, which I'm not sure about. It doesn't matter though; we've already established at least 62% of the budget tracks GDP. You can probably use the lack of growth in productivity for government tasks to argue that a given function's cost will increase linearly with GDP.

If you want to use the population + inflation metric, though, we're just returning to historical rates of growth on that scale after the drop of the mid-90s.

Jason:

Why would (or should) the education budget track GDP? Yes, I understand that salaries of individual teachers should track GDP (or inflation), but wouldn't the number of school age children be a better measure than GDP? I would expect the education budget to fall if the number of school age kids fell, rather than go up, even if indvidual teachers got a pay raise.

Of course, not even rough population numbers are a good measure. In parts of a community, the numbers of school age children may be falling (as the neighborhood ages) while newer areas of the same community may need new schools to be close to new homes being built. Building new schools is expensive and such costs cannot be fully offset by closing older schools. Still, many the demand for, and the cost of, many public services should track population, not GDP.

The cost of government should not necessarily go up with GDP. GDP goes up with productivity. For a given level of government services, as productivity rises, the cost of government services should fall. That's if productivity goes up in the government sector as fast as it does in the private sector, which of course it doesn't.

Also, as GDP goes up, poverty declines, and so should the ostensible reason for many government services. To offset this, politicians invent new needs.

I read somewhere today that something like 75 percent of young working-age French people want to work for the government. If we ever get to that point, our economy will be down the toilet.

here's the story on France:
http://www.techcentralstation.be/2051/wrapper.jsp?PID=2051-100&CID=2051-052703N

"The cost of government should not necessarily go up with GDP. GDP goes up with productivity. For a given level of government services, as productivity rises, the cost of government services should fall."

For capital-intensive government services, like, say, transport & infrastructure, yep, but not for labor-intensive services, like, say, education or policing, or healthcare. Don't see how gains in semiconductor processing will translate into smaller class sizes, or the need for fewer police.

For labor-intensive services, the costs will track overall personal income levels, which would go *up* with productivity gains.

But Tom, your argument makes no sense. In the private sector, wages rise with productivity. You're arguing that public sector wages should rise because their labor is no more productive than it was before -- that we need to pay our policement more because computer scientists got an 8% bump this year. Unless we're in danger of losing our policemen to computer science schools, this isn't true. They'll get a "raise" in the form of cheaper and more plentiful consumer goods -- but there's no reason that we should pay more for the same services while the supply of workers remains plentiful. If you'll check the waiting lists for various civil service positions, I think you'll see that we're in no danger of running out of government workers.

Anyone proposing any other metric than GDP is implicitly saying something pretty weird about the marginal propensity to consume government services ...

"For a given level of government services, as productivity rises, the cost of government services should fall. That's if productivity goes up in the government sector as fast as it does in the private sector, which of course it doesn't."

Where is John Calvin when you need him? Abstractly speaking, there is no logical reason why government spending should go up during good economic times. Human beings, however, are never entirely rational and logically consistent. They are sinners who are sometimes unable to successfully conquer temptation.

The cost of government inevitably goes up because some citizens are always eager to beg their politicians for money from the treasury to support their favorite causes. Needless to add, these elected officials are often more than willing to comply if it might increase their chances of reelection. Furthermore, it’s a lot easier to play Santa Claus when you are not the one actually picking up the tab.

PS: A few folks might be upset by my indulgence in theologucal rhetoric. Quite frankly, I could care less. It works for me.

" Abstractly speaking, there is no logical reason why government spending should go up during good economic times."

Jesus, David. Do you really think there's absolutely no reason to all to spend marginal income increases on public goods?

I should have spoken more concisely: a government program that maintains a "given service level", where that service level depends on the income, will track GDP. For example, if Social Security is defined as "pay a given fraction of the median wage to retirees", then the cost of Social Security should track GDP, as the cost is just a constant proportion of income * population, or 1/constant * GDP. Programs that chiefly pay salaries should track GDP on a related train of thought; you can't keep paying the same real salary and deliver the same service level when your employees have alternative career options.

"The cost of government should not necessarily go up with GDP. GDP goes up with productivity. For a given level of government services, as productivity rises, the cost of government services should fall. That's if productivity goes up in the government sector as fast as it does in the private sector, which of course it doesn't."

I'm not sure I follow.

Two good economy, where productivity in A increases 3% annually and productivity in B is flat; define "cost" as time expended to produce a unit, and that the cost of A & B start off equivalent.

Every year, the effort it takes to produce a unit of A should decrease 3%, while the effort it takes to produce a unit of B is flat. The "cost" of a unit of B stays the same, while the cost of A decreases.

Applied to public sector/private sector, a given volume of private goods always get cheaper, but a given volume of public goods tends not to, for various reasons. The population this year can have the same count of goods they did last year, but with a little time left over, thanks to the productivity gains in the private sector. It seems unlikely they'd spend all that additional time producing private goods; if they were so hell-fired to get more private goods, why didn't they forgo a few more public goods last year to get them? They'll buy more public goods at whatever their marginal propensity is for them.

Over the long run, the ratio of private purchases to public purchases (government spending as fraction of GDP) should approximate the relative spending desires for public and private goods.

Now, what are you arguing when you say "government spending is out of control?" That the government is ignoring the will of the people by spending more of this year's income on public goods than than people really want? Or just that government spending really went up last year?

I suspect it's the former; after all, if the public is happy with the increase in government spending, what's the big deal? A useful rephrasing is that "the government is spending more of new income for public sector goods than the public's MPC for them." Judging by the past history of the MPC, though (state spending as a fraction of GDP), the government is pretty much in line with public desires. Unless you think public opinion has massively shifted as of late, which would show up in polling, and it hasn't.

On a related note, Kevin Drum has a chart up clearly showing how closely CA state spending as a fraction of GDP has tracked the business cycle; it's bounced around 9-something percent since 1980.

Yes, California's spending tracks the business cycle, which is because they spend every extra ounce of tax revenue. They spent the bejeesus out of capital gains, and now that those capital gains have gone away, they're still spending like it's 1999. Raising taxes on vulnerable businesses during a downturn is not a winning formula.

The public goods we're getting don't seem to be adding much marginal value. State spending has gone up 84% since 1991, which is about 65% adjusted for inflation. Who thinks they're getting 65% more value from their state government than they did in 1991? Increases tend to go to things that please small, active voting blocks, such as public sector unions, rather than those that benefit the population at large. Are state residents better off because their health care workers get a 25% salary increase over 3 year for no improvement in service? If it's so reasonable, how come none of us taxpaying types can get a deal like that?

New York is a case study in public choice – as jane points out, we are getting no more “value” in terms of public goods while spending continues to go up up up.

It is the small powerful interest groups (in NY, SEIU Health Care Unions {most incredible ‘get out the vote’ operation I’ve ever seen}, Teachers, etc) who push and push spending up each and every year. They have a huge incentive to do this – the increases go directly into their pockets.

The average guy on the street does not have the incentive or the network necessary to lobby the State legislature to stop the spending.

Put down the economics textbook and open your eyes to the real world. It is not taxpayer demand that drives spending, it’s interest groups whose employment depends on it.

Jane said:
"You're arguing that public sector wages should rise because their labor is no more productive than it was before -- that we need to pay our policement more because computer scientists got an 8% bump this year. Unless we're in danger of losing our policemen to computer science schools, this isn't true. They'll get a "raise" in the form of cheaper and more plentiful consumer goods"

Yes. But they'll also get an effective "cut" in the increasing price of housing, which tracks personal income growth. So, frex, San Francisco policemen end up living 2 hours away in Tracy in the central valley, 'cos they can't afford to live in the Bay Area proper.

Jane also wrote:

"Yes, California's spending tracks the business cycle, which is because they spend every extra ounce of tax revenue. They spent the bejeesus out of capital gains, and now that those capital gains have gone away, they're still spending like it's 1999. Raising taxes on vulnerable businesses during a downturn is not a winning formula."

Or more exactly, Prop 13, with requirements for supermajorities and constraints of property tax increases, has made California exceptionally dependent on sales taxes and individual income taxes. So its revenue is structurally volatile. It's hard to find good things to say about Davis, but at least he did try, in his original proposed budget, to try to deal with some of these problems.

(Prop 13 is also the reason why you tend to see tight housing markets in CA, but gluts in commercial space).

Jane also wrote:
"State spending has gone up 84% since 1991, which is about 65% adjusted for inflation. Who thinks they're getting 65% more value from their state government than they did in 1991?"

I'm not sure about your numbers - I'm getting a 78% nominal increase 1991-2002, with a 47% real increase in real terms (using the GDP deflator) [Numbers taken from the Economic Report of the President, 2003]. Over the same period, personal consumption grew by about 43% in real terms. I'm not seeing a big discrepancy here.

“Jesus, David. Do you really think there's absolutely no reason to all to spend marginal income increases on public goods?”

Of course there are sometimes excellent reasons “to spend marginal income increases on public goods.” The problem due to the weakness of human nature, though, is that we are often easily seduced into believing that just about everything is a necessity. It’s akin to the sailor with a serious drinking problem who is about to get his check before hitting the foreign town. Rationalization is much easier when the wallet is bulging. The crap may eventually hit the fan, but in the meantime there is much liquor to consume and fast women to chase.

“Put down the economics textbook and open your eyes to the real world. It is not taxpayer demand that drives spending, it’s interest groups whose employment depends on it.”

The full time employees of Interest groups constantly petition their politicians. And as I mentioned earlier---the elected officials are often more than willing to play Santa Claus with other peoples’ money. You cannot understand human nature unless admitting that the at least metaphorical reality of Original Sin is alive and well on planet Earth.

Okay, Tom, but people still want to be police officers. Why should we pay them more if there's an ample supply of labor at the current wage? Taxes are not some pot of goodies to be divided among public sector unions; they are to be used to provide services for the public at large. If we can get people to provide those services at an adequate level for the current wage, there's no reason to raise them. In New York City, the average police officer, who has no college degree, costs the city over $130,000 a year, while the average fireman tops $150,000. That is not what is required to get someone to do the job; it's a gift to the unions from politicians to get elected.

>>The public goods we're getting don't seem to be adding much marginal value. State spending has gone up 84% since 1991, which is about 65% adjusted for inflation. Who thinks they're getting 65% more value from their state government than they did in 1991?

Do you mean spending per capita here? This comment only really makes sense if you do.

In unrelated news, a moment's consideration of the numerous ways in which a policeman can supplement his salary if he is so minded, should suffice to establish why it is a *bad* idea to pay them market-clearing rates. If someone showed up at your jeweller's shop offering to do the job of guarding your safe at night for no money at all, would you say "hurray, small cash saving for me?"

"Okay, Tom, but people still want to be police officers. Why should we pay them more if there's an ample supply of labor at the current wage?"

Because I see no benefit in public servants having their relative standard of living on a long-term decline.

To fill out my point, I'll give an anecdote. Two years ago, I was in Caracas in Venezuela. Caracas is an unbelievably unsafe place, and the poorly-paid policemen regularly take bribes. A former Miss Universe, Irene Saez* won the mayorship of the Caracas equivalent of a borough. One of the first things she did was to fire most of the police in that borough, and hire new officers & management on triple the salary. That borough is one of the few safer parts of Caracas. Sometimes the lowest price bidder ain't the best.

[Saez, until she made some poor political decisions, could have beaten Chavez for the presidency].

* You've never seen as much silicone as in Venezuela. Every middle-class woman has plastic surgery, and goes out dolled to the nines - being from San Francisco, I kept thinking I was seeing drag queens, 'cos I had a hard time believing real women would go out in public that glammed up.

Tom, that's a relevant example in any country where civil servants are commonly paid less than a living wage, given excess discretion, and are subject to less-than-transparent accountability mechanisms. But here?

Here is a comparative breakdown for a law-enforcement career in New York City, New York vs. Denver, Colorado (the area where I live):

NYC
Denver

Looks like the starting pay is comparable (depending how much you want to adjust it for cost of living), but WOOOOWEEEEE, are those some nice benefits for working in NYC. The discrepancy in paid vacation is particularly noticable, but all around the NYC officer is getting handed a pretty sweet deal.

Tom, civil servants in third world countries are paid pittances and expected to make up their salaries out of graft. Suggesting that we ought to pay what the job is worth is not the same thing as suggesting that we should cancel their salaries and invite them to take bribes instead, any more htan the fact that there are starving children in Caracas means that we need to expand the food stamp program. The comparison is silly. Our civil servants are hardly on the verge of plunging into poverty.

D^2, the population increase in the United States over ten years hasn't been that great, and high spending states like New York have experienced a net decline, even as inflation-adjusted spending grew 59%.

"Tom, civil servants in third world countries are paid pittances and expected to make up their salaries out of graft. Suggesting that we ought to pay what the job is worth."

If civil servants salaries tracked inflation rather than GDP/capita, then you'll find that public servants relative standard of living will drop. That's what you're advocating - public servants would find themselves dropping down the income distribution curve. That would have adverse effects on the quality of public services, as Dsquared pointed out above. I'd suggest we've already seem this happen in teaching.

BTW, the rule of thumb typically used when I was writing B-plans for startups was a total cost $200-250K/employee. Compared to that, I don't think the $130-150K total cost for NYC firefighters or policemen is out of whack, especially considering the additional risk (and hence additional benefit costs) they're incurring.

"D^2, the population increase in the United States over ten years hasn't been that great"

253 million in 1991 to 288 million in 2002. 14% increase.

" Are state residents better off because their health care workers get a 25% salary increase over 3 year for no improvement in service?"

So you think their provided service level would be just fine of they never got raises? They're not getting jumps like that every 3 years; there's always a long period of stagnation between 'em.

This is just a bizarre thread. I think I can sum up the complaints as "I don't like the expressed marginal preference for government spending of the public."

"Yes, California's spending tracks the business cycle, which is because they spend every extra ounce of tax revenue. They spent the bejeesus out of capital gains, and now that those capital gains have gone away, they're still spending like it's 1999. Raising taxes on vulnerable businesses during a downturn is not a winning formula."

What does this have to do with "CA spending has stayed right around 9%, implying that's the marginal preference for government spending of income increases since 1980?"

Tom, spending growth isn't well correlated with population growth. This is quibbling. If you're using a GDP deflator, cost of living cannot be simultaneously rising across America; only the cost of consuming things they didn't previously consume; most people do not spend the bulk of their budget on food or gas. Locally this can be so, but high cost coastal areas tend to be paying their civil servants far, far above the median wage for the area -- in New York, the wage/benefit package averages something near triple the median. There may be a problem with real estate prices in Silicon Valley, but that's simply not a nationwide phenomenon, and no state or municpality can afford to pay their civil servants to live in an area densely populated with millionaires -- cops in New York don't live on the Upper East Side, and we couldn't afford to pay them if they did.

The fact is that we are increasing state and local spending at a very rapid clip, even though few people have seen a corresponding increase in the quality of services provided them. Indeed, key services like education are steadily declining, despite rapid increases in spending levels. It is also true that civil service salaries are increasing at well over COLA, inflation, or GDP growth rates.

If civil servants don't like being paid what the market demands, they are free to get jobs in the private sector that pay what they feel they are worth. As someone who pays a heavy tax to support their services, I feel no altruistic urge to give them more than is necessary to obtain their services, and if you do, I invite you to start a private fund for this purpose.

Well, come to Texas. It has no COLA for state employees, no state workers' union, and rapidly declining health insurance benefits (I think that the State offers only one HMO as of now). Of course, this means that the State has dizzying turnover of nearly 20% yearly, much higher in state employee populations that service high risk populations (the mentally ill, child protective services, etc.).

Even State Auditor figures indicate that the costs of such turnover are about 4 times greater than the cost of implementing a COLA system.

As for private sector re-entry, don't worry. The State is forcing that on its own by indiscriminately cutting 12.5% of the state's FY 04 and 05 budgets from FY 02 and 03 levels, leading to up to 15,000 employees getting a pink slip starting September 2003 (up to 8-9,000 people in the People's Republic of Austin alone). Check back with us in two years. I don't think we'll be a model of lean government efficiency in delivery of services.

Sorry if I'm being provincial or too narrowly focused. I'm not even sure what the preceding paragraphs have to do with the overall arguments, other than the spend money to make money thing (like the CHIP program, where the feds reimbursement state governments with $2.50 for each $1 the state spends).

"Indeed, key services like education are steadily declining, despite rapid increases in spending levels."

Over the long run, would you expect the percentage of income spent on education to decrease, remain constant, or increase?

"Tom, spending growth isn't well correlated with population growth. This is quibbling. If you're using a GDP deflator, cost of living cannot be simultaneously rising across America; only the cost of consuming things they didn't previously consume; most people do not spend the bulk of their budget on food or gas."

I'm not following what you're getting at here, Megan. Could you rephrase it?

"Locally this can be so, but high cost coastal areas tend to be paying their civil servants far, far above the median wage for the area -- in New York, the wage/benefit package averages something near triple the median."

You're gonna have to source this, Megan.

"The fact is that we are increasing state and local spending at a very rapid clip,"

But not as a %age of GDP, as Jason said above. It seems to me you're synthesizing a crisis here to match into a preformulated argument, regardless of the statistical data.

"even though few people have seen a corresponding increase in the quality of services provided them."

That's not my argument. My argument is that for many government services, which have large labor components, wage/salary would track real increases in private-sector compensation, *for the same level of service*. The same as the hourly rate for a plumber or an auto mechanic goes up with overall wage levels, rather than with inflation, despite the fact there's been no revolution in plumber productivity.

You should also know that using a metric like productivity isn't applicable to government services, where contribution to GDP is valued at the wage/salary cost.

"Indeed, key services like education are steadily declining, despite rapid increases in spending levels. It is also true that civil service salaries are increasing at well over COLA, inflation, or GDP growth rates."

You're gonna have to source that again, Megan.

"If civil servants don't like being paid what the market demands, they are free to get jobs in the private sector that pay what they feel they are worth. "

Umm, that was my point, Megan. And Norbizness above gave an illustration of what happens when the delta between private-sector and public-sector gets too great.

"Tom, spending growth isn't well correlated with population growth. This is quibbling. If you're using a GDP deflator, cost of living cannot be simultaneously rising across America; only the cost of consuming things they didn't previously consume; most people do not spend the bulk of their budget on food or gas."

I'm not following what you're getting at here, Megan. Could you rephrase it?

"Locally this can be so, but high cost coastal areas tend to be paying their civil servants far, far above the median wage for the area -- in New York, the wage/benefit package averages something near triple the median."

You're gonna have to source this, Megan.

"The fact is that we are increasing state and local spending at a very rapid clip,"

But not as a %age of GDP, as Jason said above. It seems to me you're synthesizing a crisis here to match into a preformulated argument, regardless of the statistical data.

"even though few people have seen a corresponding increase in the quality of services provided them."

That's not my argument. My argument is that for many government services, which have large labor components, wage/salary would track real increases in private-sector compensation, *for the same level of service*. The same as the hourly rate for a plumber or an auto mechanic goes up with overall wage levels, rather than with inflation, despite the fact there's been no revolution in plumber productivity.

You should also know that using a metric like productivity isn't applicable to government services, where contribution to GDP is valued at the wage/salary cost.

"Indeed, key services like education are steadily declining, despite rapid increases in spending levels. It is also true that civil service salaries are increasing at well over COLA, inflation, or GDP growth rates."

You're gonna have to source that again, Megan.

"If civil servants don't like being paid what the market demands, they are free to get jobs in the private sector that pay what they feel they are worth. "

Umm, that was my point, Megan (and one you earlier rejected, saying that we didn't have to worry about policemen leaving to become computer geeks). And Norbizness above gave an illustration of what happens when the delta between private-sector and public-sector gets too great.

Median wage in New York City is currently in the low 40's, with benefits worth approximately $10-15K more. The average civil servant costs the city over $100K, with the police and firemen leading the pack in the high $130K and low $150K ranges, respectively. You can find the numbers yourself from the City IBO.

The point is that private sector wage increases track either the cost of living, or productivity increases -- usually both. The civil service increases here track neither. The MTA, for example, just got over 10% over 3 years, when expected inflation is negligible, the cost of living is expected to fall due to the collapse of our financial and real estate bubbles, and productivity isn't budging. This is not, incidentally, because no productivity increases are possible -- many productivity improvements could be implemented. The union doesn't want them. We may expect to see the same thing happen with the firemen's upcoming negotiation, despite the fact that utilization of the department has been falling for twenty years, and the policemen. The teachers got an enormous new package. All this, mind you, when the City is facing a multi-billion dollar deficit, and when the financial services sector, which provided probably 1/3 of the City's tax revenue in 2001, is contracting sharply. The waiting list for civil service jobs in the police department, sanitation, MTA, fire department, etc. is years long despite competitive examinations. I repeat, we are not in danger of losing people. If and when turnover, rather than retention of useless dross, becomes a problem, we should raise wages. Not before.

"Median wage in New York City is currently in the low 40's, with benefits worth approximately $10-15K more."

I'm surprised it's that low; and I think you're underestimating benefit costs. Typically, I'd think of benefits + payroll taxes as costing 0.45-0.6 times base salary/wages ($18-25K, for a $40K base).

"The average civil servant costs the city over $100K,"

Is that including non-benefit & payroll tax overhead?

"with the police and firemen leading the pack in the high $130K and low $150K ranges, respectively. You can find the numbers yourself from the City IBO."

Cheers.

Any info on the wage distribution in NY area? Frex, I'd expect (based on my own value judgements) that policemen & firefighters would be in the second quintile for skilled blue-collar workers, and teachers to be in the sixth to eighth decile for bachelor-level college-educated workers.

"Well, come to Texas. It has no COLA for state employees, no state workers' union, and rapidly declining health insurance benefits (I think that the State offers only one HMO as of now)."

I also live in Texas. We do indeed have a few problems to work out. Still, I much rather have our troubles than those of New York City. Our future looks very bright. The private sector is constantly growing and people are never endingly moving into the state. How many people are today moving into New York?

"Median wage in New York City is currently in the low 40's, with benefits worth approximately $10-15K more. The average civil servant costs the city over $100K, with the police and firemen leading the pack in the high $130K and low $150K ranges, respectively. You can find the numbers yourself from the City IBO."

Err, no I can't. I can find one document saying that using civilian versus uniformed employees in the NYPD saves $29K/employee, but nothing on median wages for city employees. You're gonna have to point me in the right direction.

>>Tom, spending growth isn't well correlated with population growth. This is quibbling.

No it isn't. It's insisting on accuracy. I've never seen you give anyone else the benefit of the doubt on using semi-attached figures, so you shouldn't ask for it yourself in your new career as a pundit.

In related news, the construction "in New York, the wage/benefit package averages something near triple the median" is as confusing as hell in the way that it merrily trips between means and medians, and in any case, the figures you provided seem to suggest that the mean New York civil service wage/benefit package is $100K, which is not triple $60K.

It is impossible to tell whether you have a point or not when you're using so many sloppy numbers, although you do rather appear to be dodging Tom's excellent point that wage increases are driven by productivity *in the economy as a whole* rather than in specific industries, so I'm assuming for the meanwhile that you don't.

Obviously, if you want to take the low end rather than high end of the range "low forties" plus "10-15K", substitute "$50K" for "$60K" above, without much loss of point.

Comments are Closed.