Below is the transcript to ShadowStats’ John Williams Explains Why It’s All Been Downhill Since 1973:
Chris Martenson: Welcome to this podcast. I am your host, Chris Martenson, and today we are talking with a man whose work I admire very much and who was single-handedly responsible for exposing me to the idea that our official statistics on inflation, unemployment, and the GDP should not be trusted blindly, ideas now enshrined in the Crash Course chapter I have entitled “Fuzzy Numbers.” His name is John Williams, and he runs the Web site ShadowStats.com. It offers consulting services to companies that need reliable data to manager their operations. What a pleasure. Welcome, John.
John Williams: Thank you for having me, Chris.
Chris Martenson: It is a great pleasure to be talking to you today. Why don’t we begin by having you explain to our audience how you came to create the Shadow Stats government statistics service.
John Williams: Sure. I have been a private consulting economist for about 30 years, and one of my early clients was a large commercial airplane manufacturer. Their primary forecasting tool was a number called revenue passenger miles, and they had a model, an econometric model, based on a number of factors, including the gross national product, which at the time was the broadest measure used for the U.S. economy. The model had stopped working, and I looked at the detail and realized there was a problem with the Gross National Product reporting. There was a problem, actually, in the trade statistics, which I corrected privately; and the revised numbers worked quite well and it did for some time until the changes in economic reporting made the GDP and GNP numbers almost worthless. As I found early on with that, I really had to concentrate on what was being reported by the government if I ever wanted to be able to give a meaningful forecast or explain what was happening. I got into the numbers with people who had been involved in the creation of the series, people who had been in the business back into the ’20s, and those that had worked for the government.
I surveyed the National Association of Business Economists for several years at their annual meetings and would ask the economists attending how they viewed the quality of government statistics. Generally, about 70 percent thought the numbers were pretty poor quality, and the interesting thing I found was in talking with people who knew specific numbers. For example, the economist for a very large retailer told me that the retail sales numbers put out by the government were virtually worthless to them. They did not have much meaning in terms of their sales. They thought they were very poor quality. He, however, thought the money supply numbers were of good quality and liked to rely on them. And so the next guy that was filling out the survey, and we would talk, was an economist for a large New York bank; and he said the money supply numbers are just of miserable quality. But I think the retail sales numbers are pretty good. And what I found was that the people who knew the ins and outs of the specific reporting knew where the flaws were, where the problems were, and in fairness to the government, it is very difficult to put together meaningful economic statistics on a timely basis; but a lot of series have been generated and carefully followed overtime.
The problem is a key series, such as the CPI inflation, the consumer price index, the employment numbers, the gross domestic product, over the decades have tended to move away from what seems to be common experience. In fact, even in the late 90s, the Kaiser Foundation and the Washington Post surveyed consumers as to what they thought the rate of inflation was, what the GDP was, the economic activity, what unemployment was. The gist of the article was, ho, ho, look how stupid the American people are. They think inflation is double what it is, and economic growth is about half what it is. But the joke there was on the surveyors because what the average guy sees and thinks is happening, generally is a lot closer to I think what most people would perceive as reality than the government reporting. And the reason the reporting has strayed so much from what at one time was something that was more in line with reality is that there have been changes to methodologies over time that tend to put upside biases in the economic reporting and downside biases into the inflation numbers. Those numbers are, again, over time, the biases are one way of manipulating the data, and that probably is a dominant form of how these numbers have been massaged over time to show happier news. You do get occasional one-shot interventions or manipulations, and it happens on both sides of the isles. Both sides of the isle, I mean back in the Johnson Administration, President Johnson was noted for regularly reviewing the, what were then, again, the GNP numbers; and if he did not like the report on the economy, he would send it back to the commerce department and keep doing so until the commerce department got it right.
In the case of first George Bush Administration, President Bush was faced with a reelection bid against Bill Clinton. He had a recession at hand. Someone from the commerce department went to a large computer manufacturer, said “Hey folks, we got to get George reelected here. Can you boost your reporting of computer sales to the Bureau of Economic Analysis?” Bureau of Economic Analysis is the entity that reports on the GNP, the GDP, and they did. And the economy, the reporting on the economy improved. President said, “Whoa, the recessions over, the economy’s booming.” People thought he was out of touch with reality, if you recall the time; and, in fact, he appeared so. And, again, here you see the average person, Main Street, U.S.A., again, has a pretty good sense of what is really going on. That is what is seen with these statistics. Probably the most egregious area of change, the one that has had the most impact on people, has been with the inflation numbers for the consumer price index.
Chris Martenson: Well, how does that really impact people? So, look, if – I am just going to play devil’s advocate – so if they are a little bit fudged, maybe this is the equivalent of a white lie – it is a harmless fib. How does this really impact people?
John Williams: It is far from harmless. From the economic side, that can affect people’s investment decisions, corporate planning. If you are using bad numbers there, you may have some shocks in terms of return that you eventually get back on your portfolio or how a sales program works. From an inflation standpoint, a lot of people have an inflation adjustment either to their wages or a cost of living adjustment to social security payments or from the standpoint of an investor. Investors generally like to beat inflation with their returns, and if you are underestimating inflation, you are not going to get an adequate return. For the individual who is seeing his salary not increase as much as you might expect otherwise, that is not staying ahead of inflation, and that is deliberate. That is deliberate, because if you look at the way the CPI was used broadly in the early automotive contracts as a cost-of-living adjustment for wages, the concept at that time was a measure to fixed basket of goods, that the CPI effectively measured the cost-of-living of maintaining a constant standard of living using a very simplistic term. They would take the price of a pound of beef, a gallon of milk, a gallon of gasoline, they would price those out and then they would price out those exact same commodities next year and however much the package had increased by, that was the rate of inflation. That is how much your income had to increase if you wanted to maintain the type of standard of living you had had the year before.
In the late ’80s/early ’90s, Alan Greenspan and Michael Boskin, who then was head of the first George Bush’s council of economic advisors, started on a program saying, “Oh, the CPI overstates inflation.” And if you would ask them why the CPI overstated inflation, their point would be well, steak gets too expensive, people tend to buy more hamburger, if they might by more hamburger, then their cost-of-living is lower. And while that might be true, that is far removed from the concept of the CPI as a measure of the cost of living of maintaining a constant standard of living; nonetheless, they pushed through a number of changes. With that, the effect is, and the stated purpose, where are you pushing this, why do you want to do this – and you will see this reiterated in the recent proposals by the deficit commission that the president put together – that the…because the CPI is overstated, if we can only reduce it adequately to a realistic level. In this case, they’re talking basically just artificially reducing the CPI. Then the cost-of-living adjustments for social security will be less, and that will help reduce the budget deficit. And what was not said, of course, that was done without anyone in congress having to do the politically impossible thing of voting against social security. And social security checks today would be about double what they are had these changes not been made.
The unemployment rate is an area that I publish an alternative measure, and I also…I publish an alternative measure on inflation, on the CPI. I try to give people an estimate of where the numbers would be if these key statistics were reported the way they were initially, at a time when it was more coincident with common experience.
Chris Martenson: Yeah, now before…John, before we move on to unemployment, I just have to ask you while I have got you here, so as I understand the Boskin Commission on inflation. They came up with three new ways that we are going to deal with these things. Substitution, which you mentioned is maybe steak becomes too expensive so we flipped to a lesser grade of meat. There is weighting, so there’s all sorts of issues around how weighting gets done. And then my favorite, I think everybody’s favorite, of course, hedonics. My question around hedonics is I get the idea underlying it, but here’s why I worry it is a one-way trip: yes it is true that the cell phone I buy this year has many more features than the one I had three years ago, but my grandmother had the same phone for over 60 years. So I see that hedonics is adding value for the new features, but it is not ever deducting – or correct me if I am wrong – for the fact that products have a much shorter lifecycle than they used to. Is that true?
John Williams: Generally, that is true; and, in fact, you rarely see – although you will see it – you will rarely see hedonics end up increasing the cost of a product. For example, if you were to apply hedonics to air travel, how would you account for the delays in all the problems in terms of security inspections now at the airport that you did not have some years ago. None of that is accounted for, and I think the concept, while fine academically, and where the academics want to play games with, the average guy does not look at it whatsoever in terms of his cost-of-living. For example, you have an additive that is put into gasoline at the direction of the federal government. This is supposed to improve air quality. The price of gasoline goes up ten cents per gallon as a result. That actually happened, and the Bureau of Labor Statistics did not account that in the CPI because it was considered a quality improvement. Well, for the guy that is looking at inflation and these inflation measures were his cost-of-living or what he wants in way of return on his investments, he is sitting there pumping gas at the gas station, he is not thinking I am paying an extra ten cents per gallon here that I am not accounting because it is going to improve air quality. No, he looks at it and he says, “Well gee, the price of gas has gone up, I do not have as much cash left over.”
And for purposes of the CPI the way most people use it, getting all this fuzzy price adjustment because people have some kind of a quasi-quality improvement that the government cannot measure. Government can measure certain quality changes, for example, you buy a candy bar and it is packaged a certain size and it says eight ounces on it. Next month it comes in the same size, but they have actually shrunk the candy bar inside, put six ounces on it. The Bureau of Labor Statistics notes data and they account for that and they change the inflation for the effect of that, and that is appropriate. It is a very quantifiable effect in terms of the price of goods and what you are getting. But when you get into the hedonics, I mean, for example, computers. That is a big area, and this is an area that is distorted the gross domestic product reporting as well, because the gross domestic product, the broadest measure of U.S. economic activity, commonly is reported adjusted for inflation. Inflation is taken out of it. And the higher the rate of inflation you use, the lower the growth you are going to end up with. The lower the rate of inflation you use, the stronger the growth will be; and amazingly not that government’s biases here are to use the lower rate of inflation, which gives you the stronger growth. Now granted computer technology is expanded significantly over the last decade, over the last year or two, people now use computers as televisions; and the government will try to adjust that into their pricing of a computer.
My argument there would be, yep, there is certainly some quality improvement, but if you are now selling a computer and television, count that as a separate product. When I buy a computer I use it as a computer. Ten years ago I bought a computer system for a thousand bucks. According to the government with all the quality changes, I should be able to buy that some computer today for 50 bucks – not a chance. I have two computers sitting on my desk, the one I bought ten years ago, one a lot more recently, also cost me a thousand bucks. I actually prefer the one from ten years ago, but I do not use all the bells and whistles. So somewhere in between what the government does with their manipulations on computers and not counting them at all probably is…there is an element of reasonableness and reality. But as it is handled, the improvements in computers are very overstated in terms of their effect on inflation in terms of reducing inflation.
Chris Martenson: Right. I have had computers for a couple decades too. A still type at the same speed I always typed at. Does not really matter to me what the clock speed is on my CPU. I do not know that I am all that much more productive. And the truth is you needed that same thousand-dollar computer to participate in the operating systems of the time, and today you still need the same thousand dollar computer, because you cannot run an old computer on new stuff. So it is a very slippery slope, but I have noticed that that slope is one-way. I have not personally yet run across an adjustment which has gone in the direction of making something appear more expensive or inflation higher. Have you?
John Williams: There have been some that the government has touted, but it depends how things are going. For example, with the textbooks for the classroom. They make an adjustment, quality adjustment, if there are colored pictures in it. If I am a student, I do not care much about the color pictures. I am looking at what it is going to cost me to buy my books for this semester. And so the whole point I am making here is that when you are looking at inflation for A) either what your cost-of-living is and maintaining a constant standard of living, or B) looking at this as an investment, what kind of a return do I need to get in order to stay ahead of inflation, you are looking at what your out-of-pocket expense here. You are not looking at the hedonics and the theoretical pleasures that you are getting from these products that the government is trying to offset against your cost of living. So what I do is…I mean, if you look at the old series, the government’s been very good in reporting a number of the changes they have made in their estimated impact on the annual rate of inflation. There is some that they have not published estimates on. One is they used to reweight the consumer price index every ten years, now they reweight it every two years, which gets you into that hamburger/steak conundrum, but they do not…it is not fully substitution-based yet. They have one that they call the chain-weighted CPI-U. It is experimental but this is the one that the budget commission is advertising as being the best inflation measure, but it is not if you are looking, again, at a constant standard of living. But if using the CPI the way it was before these games, you look at the changes made since 1990 – and I look at it pretty much on an additive basis, based on the government’s estimates of what these different changes have added or subtracted to the reported CPI inflation where right now the CPI year-over-year is showing one and a half percent inflation. If you went back to the way it was calculated in 1990, you would be up to something close to five percent; and if you went back to the way it was calculated in 1980, you would be looking at something up in the area of eight or nine percent.
Chris Martenson: Eight or nine percent. That is a full six and half to seven and a half percent higher than what we are being told it is right now.
John Williams: Right. The difference between my numbers and the government’s numbers, basically, is what is being lost in measuring the cost-of-living and maintaining a constant standard of living. That means if you are keeping up with the CPI, as a government publishes it, you are seeing a declining standard of living.
Chris Martenson: Okay. So the part that gets me as well – and I know this is really hard – I understand you cannot have a single inflation measure. It really depends. I have three children who are pre-college age, so when I look forward, I have a very different view of inflation potentially, because I am trying to factor in those college costs compared to where I was when I was 18. Or somebody who is past child rearing age and is trying to live on a fixed income and has a lot of prescription drugs. How is that…is there any method out there where people can understand better what their personal rate of inflation is? Or is a single, sort of, aggregated average rate the best we can do?
John Williams: The federal government actually does publish a couple of experimental measures aimed more at senior citizens who are younger people. They have that available on their site. It is subject to the same problems that I am talking about with the CPI-U. It is just a matter of how you weight the cost consumption. I mean, you could, for example, go through the CPI and look at the weightings there and say, oh, either my weighting in this area is such and such and you could change the weightings and calculate your own estimate. I have not done a separate work in that area, but it is doable.
Chris Martenson: Okay. So let me just take that one piece for a minute, though, because I think there is still a problem here. I think there is still a problem even if we are going to use this experimental weighting and here’s why. When I was looking at the CPI for last year, medical care services went up 3.9 percent. I think that is goods and services. And, but whenever I look at what the insurance companies are trying to claim in terms of justifying their rate increases, they are speaking to numbers that are three times higher than that. How can we have a gap that large between what companies are experiencing and reporting and seemingly justifying with lots of data to rate commissions and what the federal government is reporting through the CPI?
John Williams: Once again, depends on how the government is surveying it what assumptions at they make. I would go the practical, the real world experience in terms of what you are seeing as opposed to what the government is reporting.
Chris Martenson: Yeah. Well, I am certainly experiencing a much higher rate of inflation, and I note also that when we get to the GDP calculations and unemployment or employment numbers, you are calculating unemployment that is, I think, a little higher than twice as high as the currently reported headline numbers. Is that about right?
John Williams: That is actually…yeah, it is a little more than that. I am up around 22 percent, where the official…the headline number is at 9 percent. Again, here…what you are looking at here is a matter of definition. The federal government is pretty forthright in terms of how it measures things, you read the footnotes. In terms of their unemployment rate, it is accurate if you go along with their definitions within a certain standard margin of error. They publish six levels of unemployment, of the unemployment rate. U3, their third level, is the headline number, the one that, for January, came in at I think it was nine percent. Their broadest measure, U6, came in at 16.1 percent and that included beyond the components of the headline number, people who were marginally attached to the workforce. Those such as discouraged workers, people who meet all the qualifications of being employed or unemployed, but have not looked for work in the last four weeks, which is one of those qualifications – everything but that. And as long as they have looked for work in the last year, they are counted as a discouraged worker; 1994 and before, the discouraged workers did not have that one year time limit on them. And that is where some of the numbers start to go awry, because the discouraged workers are not counted in the labor force.
So when you see the labor force shrinking, that can reflect people who are giving up looking for work. It does not mean that they are not unemployed or that they would not take a job if one were available, in fact, quite to the contrary. They are unemployed. They consider themselves unemployed. They would take a job it they could get it. And if you went around the country, asked everyone whether or not he or she was unemployed, not only would you get pretty rapid answers, but if you totaled it up I think you would come up with something closer to my estimate, 22 percent, then what the government’s putting out there. Because primarily these are discouraged workers, because the discouraged workers and the U6 number, after a year, just fall off the government’s records. If they have been discouraged for a year, they are no longer tracked. And I estimate the longer-term discouraged workers add in on top of the U6 to come up with my estimate of 22 percent. A lot of people say, “Oh my goodness that is getting pretty close the 25 percent that was…the peak unemployment hit in The Great Depression.” Well, that is true to a certain extent, keep in mind that the unemployment rate in the peak in The Great Depression was estimated after the fact. The government did not start surveying unemployment until 1940. The government’s estimate was based on the 1930 Federal Census, early applications for social security, a variety of statistics, but not a formal surveying of unemployment. But the nature there was measuring the universe as I am looking at it, people who think they are unemployed as opposed to those who were defined as, more narrowly, as unemployed because they have looked for work in the last four weeks.
Again, beyond the last for weeks they are not in the headline number, they are not in the labor force. Yet, in The Great Depression 1933, something like 20 percent of the population was working on farms, and today the agricultural employment is less than 2 percent. And if you look at the nonfarm unemployment rate estimated for 1933, you are up around 35 percent. I think that would be closer in terms of comparison to what I am looking at, the aggregate number that is bandied about. What I am seeing right now is basically is bad as it is been in this downturn and really is bad as it has been since World War II.
Chris Martenson: So the recovery that we have seen so far in the official statistics is not showing up as much in your recombined statistics.
John Williams: No, and there we have some serious reporting problems with a variety of statistics beyond what I have discussed here that comes as fallout of the nature of this extraordinary downturn that we have been at. This is the most severe in terms of depth and length. Most severe contraction we have seen since the great depression, and the modern economic reporting system pretty much put into place after World War II, never was designed to handle a downturn of this magnitude. And what you are seeing are…in particular, are the…right now seasonal factors that are so key, and most of the monthly statistics are virtually worthless. They are heavily distorted by the economic cycle and no longer reflecting the usual cyclical patterns from holidays or whatever. I mean, they reflect that as well, but the economics is dominated. You do not have stable…you do not have stable, seasonal adjustments. So that is throwing off some of the current unemployment numbers, it is throwing off retail sales. We just had a benchmark revision to the payroll employment, which is a different series which suffered some other issues, but what that is showing, even allowing for whatever seasonal patterns you want to put in there, is that the level of employment today is below where it was ten years ago. And yet, in that period of time the U.S. population has grown by ten percent. We are not seeing an economic recovery here. This is…the downturn started at the end of 2007 officially, saw a plunging activity, through 2008 into 2009. Pretty much been bottom-bouncing sense. You have had some bounces from, sure like the economic stimulus, but we are not seeing a fundamental economic recovery. And the reason you are not seeing a fundamental economic recovery is that we have some structural issues here that simply are not being addressed.
Chris Martenson: So we’ve had a statistical recovery, which maybe is not surprising because the Fed drives by statistics, so they’ve dumped money into the right places in order to get a statistical recovery. What you’re saying is the base data is not showing us a sustained recovery; we’re not seeing the typical upturn that we would see in a post-recessionary period. Is that fair?
John Williams: Reasonably so. The thing is, even with the numbers that have shown some uptick, such as retail sales, it looks like they have topped out when you adjust for inflation. Most of the gain in retail sales last month – I say last month, that was in January reporting. Tomorrow we’ll be seeing February reporting. Most of the gain was due to inflation, and if you take inflation out of it you’re not seeing growth in retail sales. That’s an important factor because of the way that gross domestic product is calculated is net of inflation, so that the recovery is only due to inflation; that’s not a good sign. Industrial production also appears to have peaked. The housing market – housing has started to turn down anew. We’re not just looking at bottom bouncing here; we’re looking at it beginning to fall off a cliff again.
Chris Martenson: Particularly in some key markets that had thought to have been resistant; we saw that article in the New York Times yesterday, noting that Seattle has finally figured out that they had a housing bubble too.
John Williams: It’ll be called a double-dip because the National Bureau of Economic Research, which is the arbiter of business cycles in the United States, called an end to this recession back in June 2009. Then they don’t change those calls, so the renewed downturn will be a double-dip recession. In reality it is an ongoing economic disaster. When I say there’s a structural change, the problem here very simply is that the average consumer can’t make ends meet. That’s been an increasing problem over the decades, largely due to the loss of higher-paying production jobs to offshore competition.
If you go back to the 1970’s, before the 1973 recession – and if you look at the government’s own numbers on average weekly earnings, and using their inflation numbers you’ll see that the average weekly earnings – that’s on a personal basis, peaked just before the ’73, ’75 recession. Never recovered; they’re still down 10% plus. As that began to unfold, the way that households reacted explains much of the way our society’s structured today. Back in the 70’s it was much more common for one person in a household to work. Usually the husband; the wife would stay home with the kids. Today it’s much more common that both husband and wife work, maybe someone else in the family as well in order to make ends meet.
If you look at the government’s latest statistics – the poverty survey of 2009, which is the most recent release, with average and median household income adjusted for inflation (and they use a really gimmick low inflation rate with that one) – it shows that not only has household income been falling the last year or two, but it’s below its near-term peak before the 2001 recession. Household income has not recovered above that, and if you use the CPI-U – the usual inflation rate to deflate that by instead of the gimmick one – that shows that household income today is below where it was in 1973. Again, the average household has not been able to keep up here. If income growth is not keeping ahead of inflation, very simply you can’t have consumption growing faster than inflation on a sustainable basis.
You can buy some growth short-term through debt expansion, and Mr. Greenspan was aware of that back in – go back a decade and a half, two decades ago he saw what was coming here, that basically the economy was going to be grinding to a halt in real or inflation-adjusted terms, and rather than have a regular cycle and let the problems in the system have their shakeout, he basically encouraged debt expansion to keep the economy growing. You saw debt leverage build upon debt leverage, and that is what’s started to collapse in this last year or two. Right now consumer lending – consumer debt outstanding – it had a little bit of a bounce up in the last couple of months with some student loans and such, but year over year it’s still down sharply. Business loans with banks are down. Unless you have growth in debt, growth in income, you can’t have growth in consumption which is 70% of the economy.
Chris Martenson: I chase it back to – if you look at the three figures back when we had them, bless their heart, ’94, ’95 we saw both a leveling out of money growth and credit. That was the period of time where I chase all of this back to. That was a key decision; something happened there and it was the maestro, Greenspan. What he did was he opened something called sweeps, which allowed banks to effectively reduce the reserve requirements to near zero because they could take demand accounts, sweep them into non-demand money market accounts at 12:00, take a snapshot, book reserves based on that snapshot which was fictitiously low, sweep everything back in at 12:01, call it a day. It was fraudulent as far as I was concerned. They should have just come out and said, “We’ve eliminated reserve requirements.”
From there we went into a nice, tidy asset bubble in stocks, and then that blew up. Of course the Fed had to fix that, so blew it up into a housing bubble. The real question is, here we are, it seems like all the chickens have come home to roost, and it’s very unclear to me what the encore to that particular story is. They’ve tried most of the tricks I can think of. Now we’re in the period of expanding government debt; that seems to be the fallback position. We have a lender and a borrower of last resort. It’s really unclear to me, as you’re mentioning