Tuesday, November 5, 2013

Can't swing a dead cat...


So I found this site
this dramatic site that uses the Great Seal
a version of the Great Seal of the United States
with the words Slaying the Dragon of Debt overlaid on the image...

I'm looking at their About page...

"Slaying the Dragon of Debt" is a research project by the Regional Oral History Office of The Bancroft Library at the University of California, Berkeley. The first year of the project was funded under the auspices of the Shorenstein Program on Politics, Policy, and Values.

Berkeley used to be a cool place.
I'm not getting that vibe from this site.

The question of how much national debt the United States can bear is one of the most divisive issues in contemporary American politics. After reaching a postwar historic low at the end of the 1970s, the national debt has risen to a level that many observers consider unsustainable. Why has the debt grown during this time? What accounts for the federal government’s almost chronic inability to balance the budget? How worried should we be about these developments?

It is irrefutable that the national debt --
by which I presume they mean the Federal debt --
is a divisive political issue today.
It's a political issue, but it's not a political problem.
It's an economic problem. Or an economic non-problem, maybe.

Clicking their TimeLine menu item brings up a timeline graphic.
You can click different parts of the graphic to bring up a relevant page.

The first one I clicked was the Omnibus Budget Reconciliation Act of 1981
which brought me to this page, and this graph:


I didn't need to read the bold print. The flat spot that runs from World War Two to Reagan told me at a glance that I was looking at debt corrected for inflation.

Debt incorrectly corrected for inflation. You can't swing a dead cat without hitting the incorrect inflation adjustment of debt.


What's the point of "correcting" a graph for inflation? Isn't the point to show what the true number would have been, if there had been no inflation?

If inflation erodes debt, the debt burden becomes less because of inflation. So if there was no inflation, the burden would be more than it actually is.

How much more?

Suppose you borrowed $1000 at a time when the price level was 1.0. Later, when the price level was 2.0 you pay back the loan. You pay back $1000. You get off easy, because prices have doubled. $1000 buys only half what it used to buy. To make the inflation adjustment, take the 1000, divide the old price level (1.0) out of it, and multiply the new price level (2.0) into it.

The inflation-adjusted value of the $1000 you borrowed is $2000, because prices have doubled. If you only have to pay $1000 to pay off the loan, well, the other thousand dollars of value just got eroded away!

But suppose we change our mind and decide not to pay back the loan when the price level is 2.0. Instead, we borrow another $100, so now we owe a total of $1100.

Later, when prices have doubled again we pay off the loan. We pay $1100. How much of the original value has eroded away?

We could do the same calculation we did before, but with the new numbers. We could divide 1100 by the old price level (2.0) and multiply by the new price level (4.0). In fact, that's how the inflation-adjustment of debt is typically done. That's how it's done for the "Slay the Dragon" graph. But it's wrong: 2.0 is the old price level for the $100 you borrowed... 1.0 is the old price level for the $1000.

Let's work it out. Divide 1100 by the old price level 2.0, you get 550. Multiply 550 by the new price level 4.0, you get 2200. So we're saying the inflation-adjusted value of the $1100 is $2200. It makes sense at first, because prices have doubled. But wait a minute...

Prices only doubled once after you borrowed the $100. But prices doubled twice since you borrowed the $1000. You have to take the 1000, divide by 1.0, the price level when you borrowed the 1000, and multiply by 4.0 to get the right adjustment. And you have to take the 100, divide by 2.0, the price level when you borrowed the $100, and multiply by 4.0 to get the right answer for the second loan. And then you have to add the two adjusted numbers together, add $4000 and $200 together, to get the right answer.

Each year's addition to debt has to be adjusted separately, to convert from that year's price level to the "base year" price level you want to use. If you lump it all together, you get it all wrong.

In our little example, the correct answer ($4200) is a lot higher than the original amount we borrowed, $1000 plus $100 or $1100 total. The correct answer is also a lot higher than the incorrect answer of $2200.

When people use the wrong method of inflation-adjusting debt, often their answer is significantly below the correct answer. Keep that in mind when you look at my "Three Views" graph below.


I got annual numbers from FRED for the Gross Federal Debt (1939-2012) and the GDP Deflator to use as a price index (1947-2012). Uploaded to Google Drive.

First I want to duplicate the "Slay the Dragon" graph. Theirs starts at 1940 but I can only start at 1947 because of the price index numbers I got. And theirs ends around 2007, but I can go all the way to 2012. But that's okay.

I'm using the same base year they used, 2000, so that my numbers are the same or close to theirs if I do everything the same way they did. It's a good match:


Graph #2: From the sheet named "Wrong" in the Google Drive Spreadsheet

Next I show three versions of the Federal debt: the raw numbers, the incorrect, slay-the-dragon style inflation adjustment, and the correct adjustment:


Graph #3: From the sheet named "Three Views" in the Google Drive Spreadsheet

The raw, unadjusted numbers are shown in blue They start out around $257 billion, hit $5628.7 billion in 2000, and end up at 16,051 billion in 2012.

The incorrectly adjusted numbers are shown in red. This line uses the same numbers as my Graph #2. The numbers start out around $1632 billion, reach $5628.7 billion in 2000, and end up at $12,517.7 billion.

The correctly adjusted numbers are shown in yellow. They start at 1632.2, reach 9348.2 in 2000, and end up at 18,080.2 billion.

If inflation erodes debt, the debt burden becomes less because of inflation. Corrected for inflation, the burden would be more than it actually is. So, let's check our numbers. Year 2000 is the base year; let's check that year.

The raw, uncorrected numbers (blue) tell us that we had a $5628.7 billion Federal debt in the year 2000. The incorrect numbers (red) also tell us that we had a $5628.7 billion Federal debt in the year 2000, after adjusting for inflation. This cannot possibly be right, because inflation erodes debt. Corrected for inflation, the year 2000 value of debt must be more than the raw numbers -- must be more, unless there was no inflation at all since the start of the graph.

But there certainly was inflation. For many years leading up to year 2000 there was inflation. So the correct inflation-adjusted number would have to be higher than the raw number. It has to be.

The yellow line shows that correcting for inflation pushes the gross Federal debt up from $5628.7 billion to $9348.2 billion. Thankfully, we only have to pay the lower number. But if you're gonna figure the inflation-adjustment of debt, you may as well do it right.

Before you ask... No, there's not a problem with the inflation-adjustment of GDP, not this problem anyway. The problem arises because debt accumulates over many years, and over those many years prices are changing. GDP only accumulates for one year, and the change in prices over one year is small compared to a long-term change. This small change is typically taken as an average for the year. So the problem identified here does not apply to the inflation-adjustment of GDP.

So, that snazzy snap-the-dragon site makes a big stupid mess of things with their flawed inflation-adjustment of debt. They claim that "the deficit exploded under Reagan." And their graph shows it.

But their graph is wrong.


// For more on the inflation-adjustment of debt (and other "it's a stock, not a flow" variables) download my 14-page PDF from MPRA.

5 comments:

The Arthurian said...

I left them a comment this morning:

If a mistake is commonly made, it may seem not to be a mistake. It is, nonetheless.

From your TimeLine I click "1981: Omnibus Budget Reconciliation Act of 1981" and come to this page:

http://bancroft.berkeley.edu/ROHO/projects/debt/1981reconciliationact.html

On that page you show a "brillig" graph of National Debt Corrected for Inflation, and you say: "the deficit exploded under Reagan."

The conclusion regarding Reagan is incorrect, because the graph is incorrect. Specifically, the correction for inflation that is appropriate for "flows" like GDP is used for debt, which is a "stock" variable, not a flow. The result is deceiving and leads to incorrect conclusions.

I review the problem here:

http://newarthurianeconomics.blogspot.com/2013/11/cant-swing-dead-cat.html

Thank you.

Jazzbumpa said...

The raw, uncorrected numbers (blue) tell us that we had a $5628.7 billion Federal debt in the year 2000. The incorrect numbers (red) also tell us that we had a $5628.7 billion Federal debt in the year 2000, after adjusting for inflation. This cannot possibly be right, because inflation erodes debt. Corrected for inflation, the year 2000 value of debt must be more than the raw numbers -- must be more, unless there was no inflation at all since the start of the graph.

I think you've fallen into a conceptual error here, due to the frame of reference. In 2008 [or any base year] the corrected and uncorrected values would have to be the same, because there is no inflation adjustment for the instant year.

If the base year were the last year of data, what you are saying would be correct. The entire data set would be backward looking, and there would be a single reference point. As is, you're looking at the base year with the current year as reference point, and it just gets confusing.

You can always go either way with inflation. 1955 debt in 2013 $$, or 2013 debt in 1955 $$. But what value does it have, either way?

Note that in your graph 3, the red line is above the blue line until the reference year. After that, it is below. Inflation is adjusted forward and backward from the reference year, but your PoV is [or seems to be] all backward from the current year.

Back whenever, when you first got me thinking about inflation adjusted debt, I came up with your same line of reasoning - inflation adjust each flow increment and add it to the stock, then adjust that incrementally going forward, etc.

What I think now is that inflation adjusted debt is a meaningless concept. You always owe the current amount in the current dollars. What does comparing current $$ to $$ at some other time do for you? It seems to add confusion rather than clarity.

I guess the intent is to get a comparison across time. To me, it makes more sense to provide context. Debt/GDP or any other denominator value that makes sense.

At each point, numerator and denominator are in consistent $'s and the inflation problem is cancelled away.

Cheers!
JzB

Jazzbumpa said...

And it really did explode under Ron, Old Ray Gun.

http://research.stlouisfed.org/fred2/series/GFDEGDQ188S

Cheers!
JzB

The Arthurian said...

Makin me think, buddy.

"I think you've fallen into a conceptual error here, due to the frame of reference. In 2008 [or any base year] the corrected and uncorrected values would have to be the same, because there is no inflation adjustment for the instant year. "

What you say is true of deficits and GDP and flows that are summed on an annual basis. It is not true of debt, because debt includes prior-year values in the accumulation. For the base year, the base year's addition to the total would have to be the same, just as NGDP and RGDP are the same in the base year. But the prior-year values must be adjusted up, for inflation.

If inflation erodes debt, the current-dollar value of prior debt is less than the value at the time of borrowing. To compensate for inflation, the current-dollar value of prior debt must be increased, just as prior-year RGDP values are higher than NGDP values.

If you make this adjustment for each prior year, and then add the annual figures together to get accumulated debt, the adjusted accumulation is higher than the unadjusted accumulation.

Imagine a variable called CDP: Cumulative Domestic Product. We calculate CDP as the sum of annual GDP values. (This is comparable to debt as the accumulation of deficits.)

We know that RGDP runs higher than NGDP for every year before the base year. Okay, figure Real CDP as the sum of the RGDP values -- all of which are higher than the NGDP, until the base year.

The sum of the RGDP values must be greater in the base year than the sum of the NGDP values: Real CDP is greater than Nominal CDP even in the base year.

The same is true of debt.

Jazzbumpa said...

But nobody thinks in terms of CDP.

And there's a reason for that. the concept isn't especially useful.

So my question about inflation-adjusted debt is: how is it useful?

Also, I still think you're wrestling with shifting reference points.

IMO, Inflation doesn't erode debt. That's incorrect framing.

What's changing, presumably, is the ability to pay. As inflation progresses, presumably you get pay adjustments so that your standard of living doesn't fall. So you have more $$ relative to debt, or your debt/dollar goes down.

But that has broken down in recent decades, and most especially in this century. Inflation is low, but pay increases have been lower. So instead of keeping up with inflation as we did in the 70's we keep falling behind.

Where is this taking us?

JzB