Dave Deek: Okay, recording has started. How's your week been going? Kevin Hawickhorst: My week's been going pretty well. I've been scrambling to finish a submission for a magazine article, but I got it over the line. So, I've been just kicking back a bit since then to take a breather for a day. Dave Deek: That sounds fantastic. The moment between when you're done with a deadline and before the next deadline... [pause] Kevin Hawickhorst: The calm before the storm. Dave Deek: So let's get the interview started. First, you've written a very delightful article called "Eisenhower Bureaucrats." You detailed basically the government teaching a concept of process improvement called process charting and as part of this greater program called work simplification. Could you give us a bit more detail about that? Work Simplification and Process Charting Kevin Hawickhorst: So the story behind process charting was this. During World War II, the government was critically short of manpower. Almost all of the experienced people had gone to the military to be drafted or just to work in some way or another on the war effort. So there was a constant revolving door of new people, shortages of every which thing, and this led to problems with management for two reasons. First, a lot of the managers were new and needed to get brought up to speed quickly, and that wasn't happening. And second, then as now, there was bad management in the government. It became a huge problem during World War II because if you're the manager in charge of getting approvals out for contracts to build bridges, you might be the person holding up the infrastructure needed for a new war plant. So there were endless complaints about the bad management in the government. The Bureau of the Budget, which works for the president and is now called OMB, the Office of Management and Budget... Anyway, back then the Bureau of the Budget decided to do something about it. They decided that there needed to be a training program for managers. They looked around and tried to find the best approach. They ended up adapting the process that was used to train new workers in the new war industries. When they set up a new munitions factory, nobody knew how to do that work. They had to train all of them. The Bureau of the Budget said, we can take this idea and adapt it for training government managers. This was their bigger picture "work simplification" initiative, as they called it. The goal was to teach managers to think critically about the processes in their office in the government and think about how to make them a little bit simpler each and every day. They tried to get rid of red tape, to try to get rid of excessive procedural burdens, or just to find better ways of doing the same thing with more modern technology or better office organization or something like that. Dave Deek: So that's actually a very familiar story. You heard the concept of total quality management, so you also know it's a similar story especially with the way Edward Deming had described it. Kevin Hawickhorst: Mhm. Yeah. It's even more similar than you think. This didn't quite make it into the article, but the Bureau of the Budget hired management consultants to help them figure out their strategy. Deming actually was a long-term adviser to the Bureau of the Budget in the 1940s. So, some of his ideas were part of work simplification initiatives, but not the process charting itself. Let me give you a few examples. There was an agency, the General Accounting Office, the GAO. The General Accounting Office used to process each and every transaction that the government did. They would allegedly look at every single one. But this was such a mammoth undertaking that it was just a very cursory examination of the documents. They would send things back for minor typos or quibbling over the way it was filled out. But they didn't do any real review looking for fraud, and everything they did was hopelessly late because they tried to look at absolutely everything and couldn't possibly do it. So Deming's contribution to management was statistical control. He said that instead of trying to look at absolutely everything, the government should figure out a good randomized way to spot check the work of the agencies. It could then relax those controls and tell the agencies, "You need to go figure out good management yourself, but we're going to audit you on a statistical basis to see if you're managing your own work effectively or not." Subscribe To Kevin's State Capacitance The Origins of Modern Management Practices All of these ideas were in play in the government's training initiatives in the '40s. Similarly, the approach of work simplification was adapted from something called Training Within Industry. I said it was adapted from the training for factories. TWI was the factory training. It died out in the US after the war but it became popular in Japan where Toyota adapted it and called it lean. So this is all to say that the ironic thing is the government in the 1940s and '50s—their management actually looks pretty modern. It had ideas that were very much the immediate precursors of total quality management and lean, and the actual founders of these disciplines were consulting for the government. Then in the '60s and '70s they said we need to adopt more modern practices from industry. But those were the practices of the very cumbersome corporations that got destroyed by Japanese competition. So the biggest irony of all of this research is that the government was actually much more modern before it was reformed than after. Dave Deek: [pauses] You're talking about Gulick is that correct? Kevin Hawickhorst: There were a lot of people who reformed the government in a direction that was not very productive. I think the '60s were the worst period for that. Dave Deek: Okay... could you tell me more about that? Kevin Hawickhorst: This is actually what the article I wrote will be about—a significant section. But I'll give you the gist of it here. In the '60s, corporations embraced "long-run planning." That was the buzzword that they loved. They said, "We're going to have the planners around the executive, and we're going to set out the high-level goals of the organization..." Dave Deek: Mhm. Kevin Hawickhorst: "...and then we're going to have some generalist managers underneath them who translate the plan into specific actions, and then the workers on the shop floor go out and do whatever they say." The theory was this would allow them to prioritize better and have more of a long-run focus. The reality was it became much more top-down and tied down with much more red tape. So it was actually a pretty bad process. However, the US was just so dominant at the time that that wasn't clear. People said, "If General Motors and DuPont do it, then it must be modern management." Dave Deek: But there were a lot of critics, especially from the old school managers. You were talking about Deming and Juran who were both very critical of that process. Kevin Hawickhorst: Sure, they were extremely critical, but how much influence did they have at the time? They were so critical just because they thought so many corporations were making such a big mistake. The government thought that it should adopt modern practices like this. And in particular, it was pushed during the Kennedy and Johnson administrations, particularly by their defense secretary, Robert McNamara. He had been the CEO of Ford and was keen to get the government to adopt these modern management practices, as he saw them. Dave Deek: So McNamara was the guy who essentially was a snake who convinced Eve to bite the apple so to speak. Kevin Hawickhorst: He is the villain of my piece to the extent there is one. Mhm. Dave Deek: That's actually very good to know because I would have thought it was like Jack Welch in the late 70s and 80s when he was proponent of a lot of bad management practices that became so rampant in America today. I did not know it was so early in the 60s. Kevin Hawickhorst: Yeah. The rot goes kind of deep and it's especially very disappointing because management during World War II had been so good. We had this approach to progressive operational improvement, to rapid iteration, to investing in the workforce and upskilling them, and then it died completely out in industry after the war. Then people assumed it must be an outdated practice,and they actively went toget rid of it at the federal level. Dave Deek: That is interesting. Before we move on, what made you even talk about this topic in the first place? Kevin Hawickhorst: Good question. So, at the bigger picture level, the reason I care about all of this in the first place is the debate over state capacity. Particularly in the early days, you had some people say the government used to be more competent in the past. I thought okay there's a prima facie plausibility. We put a man on the moon and built the interstate highway system and won World War II and we don't do very much of that today, so it could be true. I wanted to know if the government was more effective, and I was neutral at first. What was it that it actually did to be more effective? People can talk at a high level: They hired more expert people, or they had better leadership. But that's not actionable. I wanted to know questions like: how did they write job descriptions and decide about promotions? How did they budget for their projects? And in this case, how did they train their managers? And there's surprisingly very little written about it. So that was the background for me being interested in this question. And then the actual way that I stumbled into the knowledge of this, was I was just looking through a catalog of documents. I saw something that said "the work simplification initiative of the Bureau of the Budget" and I thought, “that's one of the most boring titles I've ever heard” and I kept going on. However, two or three minutes after that,I had this thought occur to me.I said, "Wait a minute, that sounds so boring that maybe nobody has ever looked at it before, so I should go skim it to see if there's something for me to discover." And in fact, there was. It wasthis very interesting document about how the government hada brilliant approach to training managers in the 40s and 50s. Dave Deek: When I was reading it, thanks to your article, by the way. I was reading I was actually shocked in how semi-structured it is, the best way I could describe it here. It's a lot more structured than how we do things in the developing world with Scrum and waterfall. You have this weird mix between super structured waterfall and not very structured scrum so it's basically doesn't actually tell people how to think about processes or how to implement processes. I thought it was a bit refreshing on my end. Kevin Hawickhorst: Mhm. Right. That's what struck me too. There's maybe this idea that only bureaucracy can defeat bureaucracy. What I mean by that is: you have the constant complaint about government employees that they always just want to follow the process and check the boxes. Maybe that's just an inherent issue with government. So in the '40s they wrote the process and the boxes to check, but the process they wrote was "go through everything your agency does step by step and eliminate red tape." So it was a process for getting rid of process. It takes the bad instincts of government and tries to channel it in a productive direction. In my eyes it is more useful than just yelling at the government to try to be different,or bringing in a bunch of consultants to fix things up and immediately they start becoming worse as soon as they leave. Dave Deek: Despite the fact that it was management consultants of the 40s in the first place responsible for a lot of these programs. [chuckles] Kevin Hawickhorst: Sorry. What do you mean by that? Dave Deek: [clarifying] It was a bad attempt at a joke. I'm not complaining about consultants, but the consultants of the '40s were the ones responsible for work simplification, Deming's statistical controls and stuff like that. Kevin Hawickhorst: No, that's an insightful point. So even in the 40s and 50s they did bring on consultants for certain projects. It was much more limited. But for the Bureau of the Budget their consultants were often individual people brought in on a contract basis. So it wasn't like they hired the McKinsey Corporation, which existed back then even and was a consultant even then. They didn't hire McKinsey to write up their approach to management. They took the people who they thought were talented and said, "Hey, we're going to pay you to come join our staff for a year or two." And that was the status that Deming had. I think he might have been working for the census and detailed to the Bureau of the Budget, but the gist of it was they brought on individual people who they thought could contribute, rather than just hired a consultant with a vague goal and an indefinite timeline. The IRS Modernization: A Digital Transformation Success Dave Deek: Okay... let's get moving because I want to make sure that we get the hour out of the way. Let's focus on the world's only successful case of digital transformation, the Snyder cuts and the digitalization of the IRS. Kevin Hawickhorst: Mhm. Right. We're talking about very very early digitalization, like punch card computing equipment and maybe the very early vacuum tube stuff in the later years of it. Dave Deek: You take the lead. [gestures] Kevin Hawickhorst: All right. So, earlier I talked about how the federal government had rapid turnover and was losing most of its expertise. This was also true in the IRS, which then was called the Bureau of Internal Revenue, but I'm just going to call it IRS. So, the IRS not only had rapid turnover in its accountants, but they had a lot of political appointees, too, which was a very negative combination for reasons I'll explain. The way IRS used to work was very very surprising by modern standards: the regional collector of revenue used to be a political appointee. So when you paid your taxes, you would send your taxes off to a political appointee who would be like a donor or something, a party bigwig. He would look through the taxes and say, "I know for a fact that this guy actually has a lot more money that he's not reporting because I'm involved in all of the fundraising and I know all of the businessmen around here." The theory of it was that bringing on local knowledge would spur people to pay their taxes. And the reality of it is that filling your tax agency with political appointees actually led to a ton of corruption. Which was actually a very predictable consequence. Dave Deek: I mean, it's a redux of the same problem since the time of the Roman Republic. Kevin Hawickhorst: Absolutely. So I said that the wartime was a very negative combination of factors because the experienced accountants who could catch fraud were leaving, and nobody ever learned the ropes. So the revenue officials realized they could embezzle money with little chance of getting caught, which they did. And further, all of the work that the IRS did was running hopelessly late, to the point where it took more than a year to get your income tax refund. So, if you paid your taxes and were going to get a refund, you would have paid next year's taxes before you got the refund for the previous year. So it was a pretty dire state of affairs—corruption all around, which was a secret at first, and IRS running hopelessly late, which everyone knew about and complained about. All right, so that's setting the stage there, right? Dave Deek: [nods] How did we get out of this mess of corruption? Because usually once corruption sets in, it just keeps on spreading. Kevin Hawickhorst: Luckily Truman appointed a very strong Secretary of the Treasury named John Snyder, who was a businessman, I think maybe from Missouri, don't quote me on that, but longtime friend of Truman and a rather personally conservative businessman. So he got in at the Treasury Department and he surveyed their problems and concluded that there were a lot of problems, and he wanted to set things right. But he had a much more solid understanding of management than lots of people have today. What he thought was this: The IRS running hopelessly late and the corruption are two sides of the same coin, which is poor procedure and bad management, because all of these confusing procedures are what causes everything to be running late. And also, it gives cover to the corrupt officials. If nobody really understands the way things ought to work and nobody can make sense of the records and the books, then you're not going to be able to catch the corrupt people. So Snyder said, "These are two sides of the same coin. We've got to get to improving the agency and its management and we're not just going to crack the whip on corruption," although he wanted to do that too, and did. But he said, "We're not just going to throw a lot of new process and new investigations at the problem. We're going to get to the root causes of it," which again was the fact that the organization of the agency didn't make any sense, the documents it used were confusing to fill out, the process for processing tax returns on the IRS side was very outdated. So you had all of these issues. Snyder said, "We're going to try to clean up the IRS by making the IRS function better across the board." Dave Deek: That's very rare because most management I'm aware of usually will just try and implement a brand new system while letting all the systemic causes just go to fester. Kevin Hawickhorst: Mhm. Right. And so Snyder wanted to iteratively improve the existing system.This was actually one of the largest uses of work simplification in the federal government. So he started trying to get just pretty low ranking managers to start improving office procedures and making experiments with improving technology to process the returns better, at a very small trial-run scale. He and the treasury leadership started making further improvements to medium-scale issues at IRS. So, for example, they reworked the tax forms that people filled out to be much more legible and straightforward. They were previously just completely inscrutable government documents. He started combining related documents where previously people had to fill out the same information multiple times. So there was improvement for the way that the IRS worked with the public. They started trying to process claims in a more efficient way and they started streamlining the way that the office worked. What's more, he was trying to change the culture, to get people thinking about the fact that the IRS had problems and they were responsible for fixing them. He started with the small problems with the aim of quickly working up to tackling the big problems, after he had won some successes and got people invested in reform. The upshot of all of this was he laid the groundwork for big picture reforms. So when the corruption scandals really broke, Congress of course called the IRS people in and tried to rake them over the coals. But Snyder had a pretty good response to them. He said, "Shame on you. I've been working to make the IRS much better and here's my successes. It's you who've filled the IRS with political appointees who've been corrupt. And you should go fix that." And he was a little bit more collaborative than what I've said. But Congress had to admit that he made a good point because he brought to them proof that he had been fixing as much as he could. So, Congress approved his reorganization plans that got rid of most of the patronage at IRS and it also streamlined the way the office was organized, which hadn't made any sense before. Those were the tools that Snyder used to invest in further transformation. He had started by just doing pilot projects in individual offices. My article gives the exact answer but—I want to say the Cincinnati office—pioneered a lot of these things. He started working with them, starting very small, saying “All I want to do is take a single tax form and see if you can automatically add up the numbers that it contains instead of tallying numbers by hand. You're just going to do it with automated equipment and you don't even need to produce the records in punch card format. It's just a single step for a single form. We're going to see if we can automate that.” And that paid off. So he spread the word around and other local offices got to trying this out too. Eventually they said, "Okay, we're going to automate many procedures for many forms." And eventually they said, "This is working well enough that we're going to start producing the records on punch cards. We're going to start having our punch-card systems interact with the Treasury's punch-card systems in Washington." And so they built it up bit by bit. They started with as unambitious a goal as possible. They took a single task: adding stuff up on a single form. I think it was personal income tax returns in a single office in Cincinnati. When they thought they had figured that out, they started to iterate, ramp up Also, they had a long run plan. So it wasn't just a lot of individual offices flailing about, but more like a lot of controlled experiments. So his digitization initiative was... he had figured out the problems with waterfall and the benefits of agile decades in advance. [sounds impressed] Dave Deek: That is very interesting. Have you heard the concept of Shadow IT? It's basically when local offices basically buy their own IT software or they build their own custom apps or use a bunch of networked spreadsheets. So it sounds like to me that Snyder essentially... hey instead of trying to centrally control that he said "Hey this is the big picture here, you guys are allowed to have your own shadow IT services but we're going to test this stuff," right? I mean, there's a lot of differences between now and... Kevin Hawickhorst: [interrupting] That's more or less how it did work. I think that's another premise of my work is in many cases these ways of thinking and... Dave Deek: [interrupting] But I... Kevin Hawickhorst: ...these methods of management have been invented before ,and in some sense the 1940s are a better model for us than the 1990s or whatever, because that was the previous era of really radical transformation. From the paperwork era to the early computer era might be a more appropriate model for reform in the AI era than anything else. The USDA: Reform Gone Wrong Dave Deek: But let's switch from the success case to the problem child. And I'm guessing which problem child I want to bring up. Kevin Hawickhorst: I'll go ahead and bring it up. Dave Deek: The USDA. [nods] Kevin Hawickhorst: Mhm. Right. So, this really ties into what I've said was the enthusiasm for bad management ideas across the 1900s. The early US Department of Agriculture was extraordinarily good and people used to bring it up as the best department in the United States, the most competent. The organization of it was reformed several times throughout the 1900s and I think practically every reform made it worse. The classic USDA, which means USDA in the 1900s and 1910s,was composed of many agencies that each had a technical focus. So there was the Forest Service—which still exists_concentrated on forestry. There was the Bureau of Entomology which concentrated on insects. There was the Bureau of Soils which, obviously, concentrated on soils. Each of those agencies had a technical focus. They did a little bit of program administration, like giving out grants to farmers to insect proof their crops or whatever. They also would do a little bit of regulation, like condemning diseased crops. Each of these agencies had a narrow technical focus and they were able to attract pretty good people. If you're an entomologist, the Bureau of Entomology could say, "Hey, this is the only place in the United States where you're going to see such a breadth of issues in yourcareer." You can go do some regulation, you can oversee long-term research studies, you can talk to farmers about good practices, you can do a little bit of everything. Dave Deek: This is like a... [gestures for Kevin to continue] Kevin Hawickhorst: And there were lots of PhD scientists orotherwise experienced people who said, sign me up for that, right? So that was the organization of classic USDA. Moving forward, people increasingly felt that it lacked a sort of integrated focus, that each agency was off doing its own thing. They also thought that the org chart looked messy because the secretary of agriculture had always more than 10 people directly reporting to him, and it got close to 20 on a couple of occasions. What they said was, we need to reorganize the department into a smaller number of major purposes. Then it will be more thoroughly directed and coordinated and rationalized. So, I won't give the details here, but from the 20s to the 50s USDA was reorganized to take all of their science and put it in one agency over here. So now there weren't the entomologists and the soil scientists and so forth. They were taken from their original agencies, picked up, and they were all put together in the agency that does scientific research. Similarly, the administration of aid to farmers was all bundled together and put in one or two agencies in a different part of the department. And finally, regulation was combined mostly in one or two offices and it was put together in still a third spot. And then on paper, this looks much more logical. It looks much cleaner. You know what I mean? Dave Deek: [nodding] Yes, I remember this pattern of organization and I'm guessing we all know the end results, right? Kevin Hawickhorst: The end results, the org chart looked much nicer on paper, but the culture of the agencies greatly changed. It used to be that each agency had a technical focus and a mix of perspectives on it. They did a little bit of regulation, a little bit of grant administration, a little bit of research. So they had both a strong sense of mission and a mix of perspectives in there. But once you put all of the different types of experts, each in their own agency, these agencies didn't have the mix of perspectives, for one thing. So if you have the agency that's just scientific research, it was no longer working with farmers. It didn't care that much about the usefulness of its research. Its research rapidly became much less applied and it appealed only to what people cared about in universities, and not so much what was useful research in practice. Similarly, when you put all of the farmer's aid in one agency, it became completely and totally captured by the farm lobby. The agency did absolutely nothing except hand out money to the farmers. And the more money it handed out, the more powerful it was going to be. So, it was an advocate. If the farmers tell it to jump, it asks how high. So the reorganization looked nice but it got rid of the mix of perspectives and the sense of mission.The department, I think ,has really stagnated since then. USDA is a fine department today, but who would say it's the exemplar of what the government ought to be? I don't think very many people would say that. Dave Deek: Okay, so the USDA classic was pretty much organized how Apple is sort of organized now, right? You have these product lines or research lines as you like to call it where it's pretty much self-contained and they can over have redundancy or overstep on each other as long the final end product was which is regulation and agricultural security I think that USDA classic would call themselves as well but that it's constantly growing and it's constantly improving. USDA modern or new USDA, let's just call it new USDA is like any other corporation where you have all the research and development in one main organization and it's removed from everything else and it's not a problem until it becomes a problem. Is that a good comparison? Kevin Hawickhorst: Yeah, that's exactly right, I think. Dave Deek: So the same narrative of 1940s and before having figured out all a lot of this management stuff that we just sort of rediscovering now. Kevin Hawickhorst: [thoughtfully] Yeah, that's definitely the case. And in fairness to them, I will say that the government up through the 50s had many enormous problems. You'd get a skewed conception if all you do is read my work, because I'm only talking about the highlights that are very impressive by modern standards. There were many things that were completely incompetent and there were many agencies that werevisiblyincompetent, leaning corrupt. So the average member of the public or the average businessman had a reason to think that the government needed serious reform. But they threw the baby out with the bathwater. There were many management practices that they thought were outdated, but actually these practices were ahead of their time. Future Research and Conclusion Dave Deek: So let's how about this here, is there any upcoming papers I know you already talked about one of your papers, but there's any other upcoming papers that you're planning to write about? Kevin Hawickhorst: Yeah, there are a couple of experiences that I want to write about. So I'm doing two that are upcoming. One is about how Congress did oversight during the New Deal, and that's the big magazine article. Dave Deek: Mhm. Kevin Hawickhorst: And another is going to be about how agencies managed their paperwork and tried to improve the design of their paperwork and things like that. So, both of those should be out within a month or so. We'll see. So, that's what's immediately in the pipeline. But then we have this question, what other things do I want to write about? Right? So, there are a couple of buckets of things. I want to write more detailed analyses of a couple of procedural areas. So, one of them would be budgeting for projects. There was one agency, it was USDA in fact, but they had a really brilliant approach to figuring out how to fund projects in a flexible manner that still provided for oversight from Congress. So, that would be a case study in how to have Congress work productively with agencies, so that both could be satisfied that there's both flexibility and rigorous oversight. I also hope to write more about how research was done in the past because it used to be far more applied. As I said, there were these technical agencies that had ties to industries and universities and local regulators. Therefore, these agencies produced very applied research. They cared a lot about making sure they did things that were useful for the relevant parties in the American public. So I'd like to write some case studies about that, and alsopresent an overall vision that this was a desirable model of research. We should try to incorporate more of that. Dave Deek: Have you, on that topic here, have you heard there's this Dutch university and Dutch agricultural policy up until the early 2000s was actually based on what you're describing here, right? You have this massive world-class agricultural research that works heavily with the farmers and but this university also work with Dutch regulators and what made one of the biggest exporters of agricultural products in the European Union. Have you heard about that? Let me... Kevin Hawickhorst: I haven't, what's the name of that? How can I learn more about it? Dave Deek: Let me just try figure out how to pronounce it. It is Wageningen University and Research. Let me just spell this out, forgive me. It's responsible... it's one of the responsible reasons why the Dutch became such agricultural powerhouse especially as they were getting more land because the Dutch are great at getting land back from the sea, cleaning up there and turning that land into farmland mostly for cattle but also other agricultural products. Kevin Hawickhorst: I mean, good models are good models, whether it's from the past or from the present. So certainly some people have accused me of thinking that if the government did something in the 1920s or 30s, therefore I think it should have to do it that way today. which I can see where they're coming from, but I think the truth of the matter is, you want to look at success and figure out what you can learn from that. I'm very much in favor of people presenting foreign models. I just happen to think that my comparative advantage is very much the historical case studies, and that's why I personally focus on that. However, I'm equally in favor of other people writing up how other countries do things. Dave Deek: That is good to know. I think this is a great time to leave off. Does that work for you? Kevin Hawickhorst: Yep, sounds great. Dave Deek: All right, just let me stop recording. Okay. For any and all readers who made it till the end, If you enjoyed the transcript, please subscribe to Kevin’s State Capacitance