Monday, March 05, 2012
Krugman - You Decide
Krugman @ NY Times, States of Depression.
The factual errors in that article run rather deep. Let's focus on this one:
We’re talking big numbers here. If government employment under Mr. Obama had grown at Reagan-era rates, 1.3 million more Americans would be working as schoolteachers, firefighters, police officers, etc., than are currently employed in such jobs.Ah, unlike Krugman I have respect for the intelligence of the general population. You look at the data, you decide:
And once you take the effects of public spending on private employment into account, a rough estimate is that the unemployment rate would be 1.5 percentage points lower than it is, or below 7 percent — significantly better than the Reagan economy at this stage.
You have to open up this graph and take a good look at it.
The two "heavy" lines are nominal federal spending as a percent of nominal GDP (red), and nominal state and local spending as a percent of nominal GDP (purple). This are indexed to the left, so they show spending as a percent of GDP in a long time series.
The lighter lines are CPI % change over the year (yellow), Fed Funds (blue), and percent change in real GDP from a year ago. (green).
It is quite true that GDP grew far faster after the early 80s recessionary sequence than it did after our latest foray into Fixing Things All Up. However if you look at the time sequence before the early 80s recessions through approximately two years after, you see that government spending at all levels has increased relatively more in our current era. It is not government spending that is the difference.
On the other hand, if you look at the momentous drops in Fed Funds and CPI versus what we are seeing now, it is obvious that the Obama administration got dealt a much more negative set of cards than the Reagan administration. By just the end of 80s recession, Fed Funds dropped about 10%, and the rate continued to drop thereafter. That alone provided a huge expansion in consumer buying power which is not obtainable in the current era. That expansion in purchasing power (with borrowed money) combined with a state of pent-up demand and the very significant drop in CPI to multiply the effect. CPI is an important factor, and that too shows up in our current time. Note that while CPI was falling GDP was growing during the current era.
Now, if you look at the changes in government spending from the end of the prior recession in each case, it is obvious that in both cases federal spending as a percent of GDP grew at a hefty clip. But state and government local spending shrank during the sequence from the end of the 73-74 recession, and it has grown significantly from the end of the 2001 recession.
Our economy is not suffering from a deficit in state and local government spending relative to that period.
Let's shorten up the period to 74 to present and look at this in some more detail:
This graph is the same as the other, except it covers a shorter period. There is one exception - I have swapped the annual real GDP series for the quarterly real GDP series, and so I swapped to a lighter green to remind you of that.
It's a lot easier to see the relative movements in this graph, so the GDP series can be more jumpy.
Now we proceed to a little more analysis. GDP growth obviously boosts employment, but GDP has to grow considerably faster than under 2% over time to boost employment relative to population. Government expenditures are highly related to social spending, and if people don't have jobs, they don't have money, so social spending increases and they aren't paying much in the way of taxes either.
Aside from the these factors (interest rates, pent up demand, CPI) that affect ability of consumers to consume, there are other factors. Among the largest of these can be effective tax rates (lowers real incomes) and accumulation of household liabilities. In other words, if I have accumulated a mountain of debt, lower interest rates may help me pay it off more easily, but I am probably not going to go out and borrow more (nor is it likely that creditors will want to lend me more).
So let's look at those relative factors. First, household and nonprofit debt:
Here we have just real GDP for the shorter period, indexed to the left (green), and total liabilities of the household and nonprofit sector change YoY indexed to the left (red).
Indexed to the right in black as a big thick line is total liabilities of the household/nonprofit sector (nominal) divided by nominal GDP, which gives a ratio of household/nonprofit debt to GDP. 1 = debt levels at 100% of GDP.
You can confirm that the big increase in borrowing at the end of the 80s recession produced a lot of the GDP growth. But gee, we seem maxed out on our personal borrowing limits now.... So that's why the red line is hugely below where it was even before the 2007 recession hit. Back then people would lend you money and not even make you pay the interest on it. We borrowed to spend.
Okay, we borrowed until we could borrow no more. In the exodus from the early 80s recessions, you can see that we are borrowing hugely. That's how we get sustained levels of high growth. There's huge pent-up demand, and as people get jobs or recover jobs, they run out and borrow money to buy cars, first, which jacks GDP up again. So as the huge relative influence of lower CPI and lower interest rates drops, the growth of demand picks up and sustains growth.
Now the US does not have such a huge pent-up demand this time, and the private debt loads are very high. The drop in debt loads is mostly due to write-offs still. Every time a foreclosure happens and debt is extinguished, the hideous black line drops. But it doesn't produce demand. Indeed, rising foreclosures are probably reducing net demand, because now those who were living in these houses without paying mortgages and rent have to find a place to live, and in most cases they are paying much more.
It should be obvious that castigating not only the Obama administration or any possible current administration for low growth is not strongly supported. Electing anyone will not change these relative economic drags. Bad policies may increase or decrease the drags, but only slowly in comparison to the drag itself.
So let's look at the government components:
Yadayada open it up to look at it.
We've retained green GDP and black household liabilities.
We have added (blue), indexed to the left, state & local tax receipts as a percent of GDP, graphed as the YoY % change.
That's just to demonstrate that state & local taxation is always a counter-cyclical factor. You see the same in every recession except the 2001 sequence. That's because GDP didn't really drop in the 2001 sequence. The basic rule is that the taxation levels remain effectively the same, but rise as a percent of GDP. Sales taxes drop but state & local government tends to counter-compensate in other ways - one of the biggest methods is the roll in property taxes. They lag the economy because they are indexed off the growth in property values from earlier years, and in most areas, property taxes never go down. If property values decrease the local governments just increase their millage rates. Then when property values go back up they tend to keep the higher millage rate.
Red, indexed to the right, is the sum of state, local and federal spending as a percentage of GDP. This is significantly higher than at prior times, and it is one of the major problems with Krugman's thesis.
Purple, indexed to the right, is the sum of state, local and federal receipts as a percent of GDP. It ends early because we don't have state and local tax data for 2011 yet. As you can see, this percentage was also higher in 2010 than it was in the exit from the 80s recessions.
In order to prevent a governmental fiscal collapse, over the long term the purple and red lines have to average pretty close together. In both the later Clinton and Bush years, they got back to healthy levels. However it was chimera in both cases, because it was caused by bubbles.
You can see that as a percent of GDP, in our current era we are pushing more on government deficit spending to goose the economy than we were as we exited the 80s recessions. It isn't working like it did before, because there is no economic multiplier in the private economy. Krugman knows this.
So why is government employment dropping? One of the main reasons is that while state and local receipts aren't dropping as a percent of GDP, the share of state and local spending that goes to retirement benefits (both pension and medical) is rising quite rapidly. This puts intense pressure on the budgets of state and local governments.
As a percent of total employment, government employment has way outpaced private employment.
This was one of the factors that really made me blink and twitch at Krugman's thesis.
We've retained GDP, government receipts/GDP and government expenditures/GDP.
We've added, also indexed on the right in blue, the ratio of government wage and salary workers to private wage and salary workers. Over time this has tended to fall. One of the reasons is that military expenditures have fallen sharply over the decades as a percent of GDP. Relative to before the recession, government employment is a larger percent of all employment. Relative to the mid 80s, it's not very different. It's slightly lower than it was in 84-85, but not much. It is sharply increased from the ratios in the later 90s.
As of this writing, we are more than 2 and 1/2 years away from the official end of the 2007 recession. We are now seeing structural effects rather than effects of recession-induced changes.
Among those effects that we must consider over the long run with regard to their effect on GDP growth are the ratio of levels of total debt in the economy to GDP. No matter which party controls government, or any possible party that might control government, we cannot get debt servicing costs down any lower. Therefore the relative drag of increased levels of debt on GDP growth must surely increase over the next years, and if we increase relative total debt, it will surely increase more. This is the point the CBO keeps making.
We're coming to the end of this:
Everything's indexed to the left as a ratio of GDP.
The orange/yellow line is total domestic nonfinancial debt. That includes (blue) federal debt outstanding, (green) state and local debt outstanding, and everything else - corporate nonfinancial and household/non profit debt.
We still have, for comparison, total government spending (red) and total government receipts (purple). All as a ratio to GDP. Debt is the real outlier, isn't it? That's what makes this economy work differently than it did in the 80s.
Debt servicing costs plus expenditures:
Here, what I have done is retained the government expenditures/receipts as a percent of GDP, and added a series of cost indexes showing total debt servicing costs of all non-financial debt at interest rates ranging from 3.5% to 6%.
For earlier years it is not valid, because back in 74 debt servicing costs as percent of GDP were much higher because interest rates were higher. The step functions of accumulations of total nonfinancial debt derives from the fact that debt servicing costs drop so more debt can be accumulated. That ain't gonna happen no more. Short term interest rates are so far below inflation that it's hysterically funny. That, btw, destroys capital.
To get net drag, go to the right side of this graph and add the %s of GDP for either government receipts plus the debt servicing cost or government expenditures plus debt servicing cost. For example, if the average interest rate for all domestic nonfinancial debt (including mortgages, car loans, and government debt) were at 6%, (more like the mid 2000s), total debt servicing cost plus paying for current government expenditures would be over 50% of GDP. As recently as 2010 it would have been 55% of GDP.
GDP is supposed to be equivalent to national income (it isn't). If you had borrowed enough money to be paying 50% of your income on debts, that's what a BK court would be imposing on you as a result of debt restructuring and write offs. You are expected to pay 50% of your income. Only about half of those subjected to such a regime make it through the restructuring. Any bad luck, no matter how disciplined you are, prevents completing the schedule of payments.
If you pick total government expenditures plus the debt servicing, it looks much more doable, because then at the 6% debt servicing level your total payments on debts and for current expenses are only 45% of GDP (but I'm not giving you a new mortgage). However in this case debt keeps rising, and the chance that your debt servicing costs increase rises also.
You can see that we are currently experiencing slow economic growth mostly because of debt drag. We will continue to experience slow economic growth because of debt drag. Because there is no private debt multiplier at all left for more government spending, the economic response to government stimulus is very small.
As a gentle hint, when your economic stimulus package partially relies on funding companies that produce products or services that can't be sold in the private economy at even the cost of production (solar panels, the Volt, high speed rail), the net multiplier on those portions is negative. That is, for each dollar spent, the private economy shrinks over the years instead of expanding, and government revenues shrink as well.
The other portion of this are current account deficits. If you run a current account deficit either debt must increase in the private sector or debt must increase in the public sector, or the increased debt may be shared by both. In other words, if you want to hold down the increase in your national debt to let this correct slowly and less painfully, either you can buy less imports because you just buy less (drop in national living standards) or you can produce more of the goods you need to buy internally. That allows you to consume more without increasing the ratio of total domestic nonfinancial debt, which holds down your total debt servicing costs as a ratio of GDP.
The US is in a particularly bad position, relative to say Japan, because so much of the consumer goods we buy are produced externally and we run a net trade imbalance. The energy imbalance is relatively easy to correct, and it also controls how cheaply we can produce goods here, plus holds down CPI.
There are ways out. Not many, but there are ways out. All the easy ways out are gone. The US is very fortunate in that it has massive natural resources. We COULD be like a number of other countries that exploit them. We aren't, but we COULD be.
We cannot easily correct our situation by redistributing income, because unless we also fix the structural trade imbalance, it will result in our trade deficit growing.
The statements are stupid on the face of it.
For one, it assumes there can be no productivity gains in those professions.
For another, it assumes demographics are the same across generations.
By Krugman's logic, there shouldn't be an increase in the rate of employment in the health care sector or in internet infrastructure.
The man is simply incapable of thinking - he'll do enough math to back his knee-jerk reactions and then his ego turns his brain off.
I recall several cases where unions agreed to wage and pension cuts on FUTURE members in order to protect the wage and pension increases for CURRENT members. That alone will make jobs in those jurisdictions less attractive to the future job seekers.
I can recall a specific case in Chicago where jobs like police dispatch were re-classified as "citizen" jobs to save money - therefore the people in those jobs were no longer considered "police". This reclassification was also done because officers would refuse those positions as they got older since it would negatively impact their pension payouts. Any old officer transitioning to a lower-paying "desk job" like that would retire first and THEN take the job as a citizen.
This goes on at most public school districts in the Chicago area. Aging teachers "retire" to get their largest pension payout and then get hired as "consultants" (and no longer pay union dues).
There are so many games being played with these job classifications - let alone GDP and CPI - that the numbers simply can't be trusted for any meaningful data mining.
Second, much of what allows the government figures to show low inflation are the adjustments they make to the “qualify” or utility of products. So the fact that a car has better life or is safer of computers work far for efficiency is carefully calculated by the central committee to show a lesser inflation, after all you are getting more. But this excludes the fact that to get the product you must still spend more money, unless you buy used. In the case of computers that is generally not a good idea. In the case of cars, substitution of used cars was a viable alternative, until the government paid people to destroy perfectly good cars (cash for clunkers) which raised the price of used vehicles to the point where they are not that good a deal in some cases, unless you cannot take on more debt. In those cases clunkers are the only option My Mustang sports care from 1988 was 10,500 dollars, out the door. By government figures, if we apply their CPI from 1988 to 2010 it should have cost me ~19,100 dollars today (source http://www.westegg.com/inflation/infl.cgi ). Yet the equivalent model today, which is of course a much better car, costs me ~37,500 dollars. 19,100 dollars will not buy any kind of a good sports car. In essence for an entire class of durable goods the government is completely lying about costs by saying though one spends more money, the product you are getting is so much better is actually less costly. This is a blatant fib
One last comment, if you look at the CPI figures for the US from 1998 to the present day, the run up in housing from the late 1990s to 2006 and the subsequent decline since 2008 seem to have had no effect on the CPI, as the government is using the “equivalent rent” model for housing costs. Yet the reality was that housing costs were the biggest single expense for most people who bought in the 1998 to the present day, the debt service to have a nice house being what people spent their money on. The fact that he Government model somehow does not factor this in is ridiculous; folks cannot buy a house or service the mortgage with an “equivalent rent” model.
The final effect of all of these “adjustments” is that the economy does not seem to be growing at all in middle class America, quite the opposite; we seem to be in an extended recession, with small but noticeable declining purchasing power over the last 5 years.
So the fact is one cannot really figure anything from government figures, they are quite deceptive.
At some point, I noticed the Mr. Krugman would take some data and draw a conclusion from it, then at some later time take precisely the same data and draw a completely opposite conclusion from it. The only changed variable that I could find was usually a change of talking points on the part of the Democrat party. On one occasion, he did this on consecutive weeks--one week he came to one conclusion (which was congruent with the Democrat party's stance), then the following week he did a 180 after the Democrats changed their view just in time for the Sunday chat shows.
Since that week, I haven't cared a whit about anything Mr. Krugman writes.
It's hard to compare much from today to the conditions in 1983. Our demographics alone are so different that it changes everything.
I think Krugman is writing to comfort himself, and sustain an illusion of infinite ability to manipulate.
Rural households that spend much more on transport relative to the nation are seeing much higher CPI.
Because I try to estimate money flows, I concentrate on spending for the lower value commodities (that gives a floor) in each category.
But it cuts both ways. At times when a lot of the population is chasing the dream (higher asset prices, quick large profits) inflation for less necessary items will sometimes rise much faster than for the regular items.
Higher education is a good example of that currently.
Links to this post: