The Canadian Bitcoiners Podcast - Bitcoin News With a Canadian Spin - US Economic Data Collection - CBP Quick Currents

Starting point is 00:00:00 Friends and enemies, welcome back to another quick current. We're going to talk a bit about what I think is an important topic these days as data collection in the U.S. kind of comes into focus for a lot of people, given the criticisms of the president about the BLS survey. Bitcoiners obviously take issue with the inflation data and its collection methodology. I've mentioned on the show that, you know, in the States,

Starting point is 00:00:23 before COVID, about 10% of the CPI data was gathered via sort of inference. So you might decide that survey respondents are not providing enough data or not responding enough. And instead of using the items you like or the data points you usually collect, you go to adjacent items. And, you know, 10% is pretty high for CPI, given that, you know, trillions of dollars trade on that data on release day and leading up to it and following it. Obviously, it's huge, huge data point. Since 2020, though, that number has jumped up to 30%. So 30% of the CPI data, arguably the most important metric when you talk about metrics that impact both retail and institutional level financing and investing, it's obviously that there's some potential for political outcomes there and political meandering around the margins, let's say. But we should talk about the data collection methods here.

Starting point is 00:01:24 So I'll warn you, I put this together when I was kind of half on painting. killers last week off my ACL surgery. But I still think it's pretty good. So we're not going to read the, we're not going to read it exactly like I normally would, but we'll go through it together. First, so, sponsors, obviously, easy DNS, lead sponsor. Best place for you to host your content. Mark is a friend of the show, friend of mine, friend of lens. You can go there and do a lot of stuff if you're Bitcoin or obviously virtual private server options for nodes and Noster Relays. But the big thing for Mark is that it's an easy service to use. You get great support from Mark. obviously, and then for a couple extra bucks, domain share, easy mail, stuff like that,

Starting point is 00:02:00 really going to protect your data. We're hearing more and more about data thievery. And as we talked about on the other podcast, Accessive Easy, there's always something new in data breaching and overreaching, and Mark protects you from both. So go there, use our promo code CBP media, get half off your first round of buys, bull Bitcoin and the second sponsor. Use our code CBP and get a quarter point off of your fees for life.

Starting point is 00:02:23 And you know, you're using them not because the fees are cheap, they are. use them because they're not custodial. They don't rug you. They don't shock on KYC you. You can pay your bills there on their bills platform. I'm not just talking gift cards. I'm talking mortgage and credit cards. And then obviously we love the team.

Starting point is 00:02:39 When we look for sponsors, we always try and align with people that we think are mission-focused. And I think we've really done that with these guys. And so I encourage everyone to be using bold Bitcoin for their buys. And it's, you know, we got a little bit of a dip here. It's 125 down to, you know, 107, 108, whatever we're at now. Imagine saying that four years ago. Anyway, get in there.

Starting point is 00:02:58 Start buying them. 256 heat, our friend, Twan, putting together Bitcoin heaters for the home. It's great. Get paid to heat your house, man. I have one running right now behind me. You can't even hear it. And it's been hashing away for a few weeks now. In terms of efficiency, you're not going to get much better than this.

Starting point is 00:03:15 Control it from your phone. Have an ambient heat measurement at the outlet, whatever you like. This thing responds in kind. These are S-19s. And, you know, whatever your application, if you're, you're heating a, you know, plane hanger or a man cave, or in my case, maybe in September, a garage gym, you're not going to find a lot of better options than this. So 256 heat, just go over, tell them we sent you. And I hope you enjoy it, man. You should be mining. Decentralized mining is

Starting point is 00:03:41 most important thing about Bitcoin. And then finally, the Bitcoin conference, 2025. In October, I am the master of ceremonies. I'll be looking forward to seeing all of you guys there. We don't have a promo code yet, but you should be attending this conference. I'm biased, and I'm part of the selection sort of process and committee for panel, panel choices and whatnot. But I'm really proud of what these guys have done. I'm proud to be affiliated with them. And I think, you know, for the first time now, Canadians have an option at home for a conference that is free of slop, free of clickbait, free of shit coins, and really does provide a focused message that's good for both the newbie and the suit. And that's a difficult balance to strike. But I think Dan and his wife, Manuela, have done a

Starting point is 00:04:22 great job striking that balance. So definitely go over, check out the conference. Just Google a Canadian Bitcoin conference. Like I said, it's in October this year. I look forward to seeing you guys there, man. So let's have some fun together. Okay, so first we're going to talk about a little bit of this data collection here. Systematic data, economic data, specifically began in the early 19th century. The 1810 decennial census included questions on manufacturing, which was the first time in the States anyway, a federal effort to quantify economic activity beyond just population counts really took place. And this evolved later into the census of manufacturers, which by the late 1800s provided a more detailed snapshot of industrial

Starting point is 00:05:03 output. And then the modern era of this collection, I think you could say, kicked off the early 20th century in 1905. The economic census was the first conducted primarily by mail. So it's this sort of survey method you still see use today in a lot of ways, though not always by letter an envelope. And what this did was pull firm lists together from other federal agencies and administrative records along with other means as well. And this shift from door to door to mailed really improved efficiency, but obviously there was new challenges. We've talked about a few of them and I mentioned one in the intro. And the big one obviously is non-response, but we'll get to some of this stuff a little later on. After World War II, the system exploded in terms of

Starting point is 00:05:44 sophistication, the Employment Act of 1946 mandated that the government promote maximum employment, well, we've heard that a bunch times in the last few years, along with production and the maintenance of purchasing power, which needs really good data. You can't measure those things unless you have really good data. So agencies like the BEA, the Bureau of Economic Analysis, were established in 72, 1972, that is, under the Department of Commerce. And they started taking center stage for national accounts. The BLS was founded in 1884, but modernized in the 30s

Starting point is 00:06:21 during the Great Depression. And they began producing unemployment figures. So you're starting to see this sort of puzzle of data come together. And then obviously the Federal Reserve in 1913 has compiled, you know, and since 1913, to be more clear, has compiled banking and monetary data,

Starting point is 00:06:37 which, you know, with tools like the Fred, Federal Reserve Economic Data, you can go on that website, by the way. It's available to everybody. With Fred, Archive, historical series since the mid-20th century, you know, that data is available for everybody. And then today, the U.S. has 13 principal stat agencies from Energy Information Administration, the EIA, tracking oil inventories to the USDA's National Agricultural Statistics Service,

Starting point is 00:07:03 monitoring crop yields. Like, there's tons of data out there. And so you've got to ask yourself, how does all this really work and how does it work together? At the core of all this, the U.S. Economic Data Collection methodology blends, surveys, administrative records, and nowadays you're starting to see more big data sources lumped in there as well. If you look at the BLS's monthly employment situation report, which you might know as the jobs report, it draws from two surveys, which is the current employment statistics, the CES survey, which polls about 142,000 businesses and government agencies for their payroll data, and the current population surveys, the CPS, which is a household survey of about 60,

Starting point is 00:07:41 thousand homes for unemployment rates. The CES provides what the government would say is establishment level data on jobs, hours, and earnings, while the CPS captures labor force participation, including self-employed workers that get missed in payroll data. So that's good. Over at the BEA, the GDP is compiled quarterly from thousands of sources, Census Bureau retail sales data, IRS tax records for corporate profits. I've got to be missing some here, but these are the important ones. At least this is what I thought earlier in the week. Yeah, Census Bureau retail sales data,

Starting point is 00:08:16 IRS tax record for corporate profits. And then also, you know, tangential stuff like satellite imagery for construction estimates, but you didn't know that. The advanced GDP estimate comes 30 days before quarter end or after quarter end, I should say, followed by the now famous revisions as more data rolls in. The Census Bureau's Economic Census, which has conducted every five years,

Starting point is 00:08:40 surveys 4 million businesses for detailed industry statistics, and they use stratified sampling to ensure that there's quality representation across different sizes and sectors, which is key, obviously, in an economy that's as diverse as the United States. For the trade data, the census and BEA collaborate on monthly exports, imports, record from customs, records, and surveys. So there's a lot of data going into this stuff. And then there's niches that get filled by other federal agencies,

Starting point is 00:09:10 So the Federal Reserve's beige book aggregates anecdotal data. That, to me, is a mishmash that you should never see anecdotal data from regional banks, while the EIA uses mandatory reporting from energy firms for weekly petroleum stocks. And these methods often involve seasonal adjustments, which is you would just call that like smoothing out holiday hiring spikes, for example, in retail data, and then benchmarking to comprehensive censuses. More and more, you're seeing agencies incorporate alternative data. So the BLS, for example, experiments with credit card transactions for CPI, and the BEA uses satellite data for economic activity in remote areas.

Starting point is 00:09:50 It's decentralized, but it's coordinated by the Office of Management and Budgets. Statistical Policy Directives branch to ensure that there's confidentiality and no interference, though, obviously. There's some questions about that these days. here's where things get tricky though there's issues with the data and methodology and no system is going to be perfect but this one presents some significant challenges the size right is obviously a problem

Starting point is 00:10:16 but we'll get to that the first thing that I want to talk what is revisions this is the thing that got the BLS chief fired two weeks ago whenever that was the data isn't set in stone when it's reported the initial releases are closer to being estimates than you know

Starting point is 00:10:31 coming down on tablets from Moses Sort of the best example of this, obviously, recent examples in August 2024, the Bureau Labor Statistics revised down job growth for the prior year by almost a million jobs, 800K, which was the largest adjustment since 2009. And, you know, during an election cycle revealed that the labor market wasn't really as strong as everyone thought. And so the question that most people had was why. And the CES survey's initial sample misses a few firms. So annual benchmarking to unemployment insurance records, you know, should correct this. And, you know,

Starting point is 00:11:11 sort of in the same vein, GDP revisions can swing by billions over the course of a year. The BEA's 2023 quarter four advance estimate was revised by almost half a percentage point after better data came through. So these are significant numbers, especially when you're talking about equity markets and bond markets moving on these things, not to mention, you know, political whims. And election strength for the party in power and the challenging party too, especially over there. There's also like sort of method flaws, methodological flaws that would amplify these problems. So I mentioned earlier declining response rates. They're down to 50 or 60 percent in some surveys.

Starting point is 00:11:51 And then that's clearly going to introduce bias. And, you know, non-respondents are going to differ significantly from participants, which is a problem for the data. The BLS is 2025 cuts to data collection for the conservative. price index, which they said was due to budget constraints, reduced outlet sampling, potentially causing significant variance in inflation measures, among other things. And the seasonal adjustments using models like the X-13 Arima seats, S-E-A-T, that's a acronym I didn't know before, can go over or under and correct kind of violently from time

Starting point is 00:12:27 of time. During COVID, for example, they misfired leading to super volatile unemployment prints, which, you know, during COVID, there was a lot of policy decisions made on the back of this data. And the data wasn't good. It just wasn't good. There's also problems with sampling errors in household surveys like the CPS, which, you know, has margins of error close to 0.2% for unemployment, but even higher for certain subgroups. The Census Bureau's American Community Survey, which feeds into poverty stats, really struggles with undercounting, immigrants, obviously, and methodological changes on the table for these things. Well, in the BLS's 1994 CPI overhaul, there was an attempt to account for substitution bias, which you should know

Starting point is 00:13:10 what that means, but it's basically shopper switching from beef to chicken, right? This is the sort of commonly referred to example. It was controversial, and they were accused of understating inflation to cut social security costs. We've heard this before. So these arguments that, you know, inflation data is just there to preserve affordability for entitlements. not the first time. The EIA's oil inventory data has faced scrutiny for discrepancies for a long time with private trackers like API, sometimes due to classification errors and refinery inputs. That's a bit over my head, but, you know, we talked a bit about stuff like, you know, how many barrels are being produced and, you know,

Starting point is 00:13:45 these things are important for economic growth numbers, right? If oil is not being produced and presumably there's no demand, but if they're getting these inputs wrong, then, you know, it's hard to really draw conclusions from them. There's other accuracy problems, too. and, you know, one of them clearly is the potential politicization of the data. And so in 25, you know, just a few weeks ago, Trump fired the BLS Commissioner over a weak jobs report. And that caused a lot of fear of data manipulation. Economists warn that weakening agencies can erode trust. That's true.

Starting point is 00:14:20 And make data more volatile and prone to larger revisions. Also true, if you're trying to juice numbers, then naturally the revisions are going to be more volatile. The Fed is griped about revisions complicating rate decisions, 2024's jobs, data, swings, delayed cuts, obviously. But it's interesting to me when Powell gets on stage and people ask him what he's looking at, he always says he's data driven. But there is an understanding at the Federal Reserve that the data is bad. So as I mentioned on the show, you know, there is no data that is worth putting a fulcrum on

Starting point is 00:14:50 in terms of policy rate decisions for sure. But among other things, we have to consider that the data we're getting weird, the data that the United States is getting is bad. there's other concerns too that are not methodology related privacy is a big one 86% of Americans worry about the data security worry about data security more than the economy itself agencies like the census pledge confidentiality for example but there's been breaches and so people don't respond there's also big data fears now this is obviously a more modern problem but there's a ton of big data integration the BEA's use of

Starting point is 00:15:24 private data sets could expose personal info if not anonymized via something called differential privacy techniques. I didn't know what that was before, but basically it's trying to anonymize survey data. People don't believe that their data is anonymized. I know, I'm in that group. A Pew survey found that 81% of people feel like they lack control over company-collected data,

Starting point is 00:15:43 which extends all the way to government. And that distrust causes people not to participate. 29% of privacy-worry households don't fill out online polls. Like, it's crazy. No one has a phone, and then you also don't fill out the online poll. you're basically under the radar as far as the data is concerned. This is going to be a problem as far as demographics go to.

Starting point is 00:16:03 I don't have that in the notes here. But, you know, just thinking about it now, nobody my age or younger is answering the phone or doing an online survey. So where's the data going to come from? It's an interesting question. Political influence, obviously, another red flag. We'll talk about 2025 again here. Trump threatening to, you know,

Starting point is 00:16:19 having quotes here, fix the data. I know he doesn't mean, like, fix like the fix is in, but he means fix like remedy of the problem. but it's it's unfavorable as far as perception and people are pointing to the Nixon era influence on the Bureau of Labor Statistics. Howard Lutnik's nomination is the Commerce Secretary, raised alarms about the manipulation of data as well. And economists, you know, this sort of nebulous group again, fear that could mirror authoritarian tweaks, you know, undermine markets and things like this. Investors rely on impartial stats for Fed

Starting point is 00:16:51 predictions and the Fed relies on impartial stats too, but it doesn't seem like a lot of people believe they're getting them rightly or wrongly. And then privacy laws, like potential federal mandates, could stifle innovation. So GDPR-style rules in Europe cut firm investments, which is a sort of a cautionary tale for U.S. policymakers. The timeless, timeliness, I should say, versus the accuracy of the data is another big tension. Flash estimates, fuel markets, but invite errors I have here, which is true. I mean, everyone likes a good hot take. We point the CPI's volatility in 2025 for that. That was caused by reduced sampling, as I mentioned.

Starting point is 00:17:31 And then, of course, there's equity issues. I don't put a lot of stake in this, but it's, you know, I'd be remiss if I didn't mention it. Data often underrepresents minorities, skews to poverty or unemployment, and it just keeps policy biases going, which is a problem. You know, overall, I just say that U.S. economic data collection is very difficult to do. It's a marvel of different systems working together over the course of close to 250 years at this point to try and get as much data as they can and then provide

Starting point is 00:18:05 quality policy outputs. But it's difficult, obviously. And it's fragile now. Probably more than this ever been, not just because of the Trump stuff, but because people just don't care about surveys. They're not answering their phones. They're not going online. They're not doing any of this stuff. Agencies like the BLS, BEA, and census also take a lot of federal money. And there's you know, a lot of, there's a significant lean now into cutting programs. And those programs, I think at some level have to be maintained. I don't know at what level, but they can't just be axed. Because without that, you have, you know, what's basically a $30 trillion economy not looking at any data at all, good, bad, or otherwise. So, you know, all this is interesting

Starting point is 00:18:49 stuff. And Bitcoiners think about this quite a bit. And I hope you guys are thinking about it too. So take that with you and I hope you enjoyed this episode. We will see you next time.

The Canadian Bitcoiners Podcast - Bitcoin News With a Canadian Spin - US Economic Data Collection - CBP Quick Currents

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.