Chicago Booth Review Podcast Do Better Accountants Improve Productivity?

Mike Minnis: Again, accountants have studied the way accounting reporting information changes behavior all the time. You think about earnings management and the way managers try to hit earnings per share targets and so forth.

Hal Weitzman: Do accountants merely measure company activity or do they actually make it better? Welcome to the Chicago Booth Review Podcast, where we bring you groundbreaking academic research in a clear and straightforward way. I'm Hal Weitzman. Today, I'm talking with Chicago Booth's Mike Minnis about his research on auditing and productivity. We're trying to get to the heart of an economic chicken and egg issue. Do audits make companies perform better or do effective managers choose to get audited? Mike Minis, welcome to the Chicago Booth Review Podcast.

Mike Minnis: Thanks for having me.

Hal Weitzman: We're delighted to have you to talk about financial reporting and productivity. Maybe two things we don't often put together or don't think of as a causal relationship. Here's the puzzle that we start with, I guess, is that companies in the 90th percentile of productivity produce twice as much as those in the 10th percentile, the least productive, using the same inputs. But economists have kind of had a hard time explaining that. Why?

Mike Minnis: Well, so just to kind of set the stage what we're talking about here, is when we talk about productivity it's level of output for given levels of input. We've got lots of faculty here at Booth that have looked at this considerably and trying to explain why this could be the case. And I think it's multifold as to why this is the case. One of them is kind of Econ 101 would say that we shouldn't see this because competition is going to force all these firms towards the middle. Good firms are going to get competed away. Bad firms are going to get flushed out of the system. And yet persistently we see this distribution difference.

One of the reasons that we came into it as accounting researchers was the idea of measurement. We have to remember that these distributions that we as economists or empiricists look at are the result of data and databases, which are there because accounting people put them there through the activity that firms conducted. And so what's kind of crucial to this process is not just the fact that some firms could be more productive, but that measurement may also play a role. And that's kind of where we stepped in. Could we kind of help think about the extent to which measurement of this economic activity could matter to that distribution? And I think a lot of the frictions that go into it is that, and the things that we've thought about prior to our paper that went into this are things like some managers choose better practices than other managers.

They hire and fire better. They control their inventory better, all of those types of practices that they do. And it kind of struck me that this is also what accountants help managers decide to do. They help measure inventory. They help measure the productivity of their employees inside the company. And so our kind of starting off point was maybe accounting both because it's a measurement and the data that goes into the data sets where we're measuring this productivity matters. But it could also matter because accountants inside the firm help managers make better decisions. It's better information leading to better productivity.

Hal Weitzman: Yeah. And it's fascinating because as you say, we know on the one hand that measuring stuff changes behavior. Across all sorts of disciplines, we know that. And yet we think of things like accounting as just passively collecting information. As you say, and you actually call it an information production technology.

Mike Minnis: Yeah, fancy phrase there. That's right.

Hal Weitzman: Which basically means it changes behavior, right?

Mike Minnis: It changes behavior. Again, accountants have studied the way accounting reporting information changes behavior all the time. You think about earnings management and the way managers try to hit earnings per share targets and so forth. But we often think of it as just a passive activity. It's the referee, I guess, if you will, on the playing field, but doesn't actually change the course of play. Let me step back for one second. Before academia, I was in practice where I did a number of those in corporate finance, I did public accounting, and then was a fractional CFO. And one of the things that I saw in practice whenever I'd get hired was an accounting system was in a mess. We tried to help them make it better. And then two or three months down the road, the managers would say, "How did you know this? How did you know that we could do this differently?"

I said, "It was in your accounting system. If you just would've taken it seriously, you could have seen how profitable certain customers were, how productive certain divisions were." And that caused them to think about reallocating capital. How do we choose different inventory levels? Because they had that information. I saw that firsthand how good accountants could come in and help a firm make better decisions. So it wasn't just, yes, we had to do it to check a box for a tax return to get it completed. But it actually led to better information for more strategic decision-making purposes.

Hal Weitzman: Right. And actually you are finding that, as you said at the beginning, that this is leading to more greater productivity. In fact, reporting quality, according to your research, explains up to a fifth of the productivity gap that we described at the beginning. So just to give us a sense of the scale, if accounting is important, how important is it? How does that compare to other things like IT or human capital that we more traditionally think of when it comes to productivity?

Mike Minnis: Yeah, good question. One of the things I'd say is, let me recap what I think the headline finding is of this particular paper, is that companies who have audited gap financial statements have higher levels of reported productivity relative to those that don't have audited gap. Now, a key question that we always try to think of as the causal question, does this accounting actually cause that productivity difference? And we can talk about a few things that we try to get to in our paper to kind of pull that apart. I will say that that's still an open question in my mind. The headline finding is we find that those two things kind of go together.

Hal Weitzman: But It could just be the better run companies have better accounting practices.

Mike Minnis: Exactly. So the start of this paper came from, I've known about this finding for years. This is a good example. I talked to PhD students about you have a finding, but I was never satisfied with this finding that I knew what I was actually finding. This productivity related to whether or not they produced audited gap financial statements. And at some point you have to say, "I've got this finding. Do I stop because it's not perfect? Or do I at least go forward with it because I think it's interesting?" And we kind of went forward with it because I thought it was interesting, but the following things still bothered me. And that is there's a very well-cited part of the literature that looks at this 1090 distribution spread, and they've collected it's Nick Bloom and John Van Reenen have this paper where they went out and surveyed managers and said, "How do you hire people? How do you fire people? How do you manage your inventory?"

So they collected this all and created a score, and it's a management practices score. And showed that it explained approximately 10 to 20% of that 90 / 10 distribution spread. Meaning you look at the top quintile of those performers on that management practices versus the bottom quintile. This top quintile are about 10 to 20% better, roughly speaking, kind of ballpark figures. But it was funny, on our audit set of findings, I saw Nick Bloom in a conference and I said, "Nick, what if I told you just one question has that same economic magnitude? And that one question is, do you prepare audited gap financial statements?" And I think he was intrigued by one question also having the same. They did all of this effort to get all of these questions and with our one dimension had a very similar economic magnitude.

But the one key question we had in our back of our minds is what if we were just simply finding the same sets of firms that Bloom and Van Reenen found? In other words, just to your point, Hal, those firms that got audited gap financial statements were also doing everything else well along dimensions of hiring and firing and managing the inventory and all the other things that they surveyed. And fortunately, Nick was also curious about this. And so we had a working paper that was using tax return data and some other sets of data that did this finding. But we got brought onto what we call the MOPS team of the US Census, the Managerial and Organizational Practice Survey. Every five years, the census does this manager practice survey. And they ask all these questions and we got to put on one or two questions about accounting.

And so now we had a data set that had both the management practices questions and the audit questions, gap accounting types of questions, and we could do what we call a horse race. They answered both sets of questions, which one seems to matter? And that was very exciting. We got this done. And if anybody's worked with census data and trying to get the survey done, took a long time to get it cleared. And then COVID hit. And then we had to delay by one more year to find the answer to this question. Finally, we got the survey out, the survey comes back in, and I had a research assistant work at the Chicago Fed in this confidential data set all somewhere in a windowless room to clean and prep the data. And I met her there one day and I told her this story about years we've been wondering about the answer to this question.

So before you hit the go button on this analysis, I wanted you to know how long it's taken for me to think about the answer to this question. And the question we were asking was, if you kind of include both of these explanatory variables in the same regression, does the audited result kind of go away? In other words, it's just good managers doing this practice like all the other practices that they were doing. And so we ran the regression and the first step we did was we replicated the management result. It's there in the new set of data as well. We kind of replicated by itself our audited variable that was very similar in terms of the results that we had with our tax return data that we had before. And then we ran the magical regression where we put both of them in there.

And it's one of those... We were talking about this before the podcast recording, it's like one of those kind of magical moments in research where you sit there and you see this result of both variables coming out, where what happens is what exactly you'd expect in the sense that if they kind of work together, there's slight attenuation on both variables. In other words, they kind of are correlated. In fact, they're very correlated in terms of the good managers are also getting financial statement audits, but they both have explanatory power. But a cool result is when you interact them, it amplifies the result. In other words, good managers who do all these other practices in the same right ways, who also have better information are yet even more productive. There's a multiplicative effect here on that interaction. So you pair good practices along a lot of HR dimensions and inventory and operation dimensions with good accounting practices. And what we find is those are the most productive firms in the economy.

Hal Weitzman: In other words, the good managers without the good accounting data are less productive.

Mike Minnis: They're less productive than good managers with better data. Absolutely. Now, back to the causal thing, and this is where our paper stands now is that set of results. It still doesn't eliminate the fact that they could be just better managers on other dimensions that are unmeasured. But it certainly gave us the confidence to go forward with a paper that says, "Look, there's some really intriguing results here. And certainly we can control for all these other dimensions of which managers are performing." We're not causally treating, but now we start to get into some very cross-sectional tests that I'm sure we'll want to talk about next, as to how we're trying to tease out what the potential channels or mechanisms could be for this set of results.

Hal Weitzman: If you're enjoying this podcast, there's another University of Chicago Podcast Network Show that you should check out. It's called Nine Questions. Join Professor Eric Oliver as he poses the nine most essential questions for knowing yourself to some of humanity's wisest, the most interesting people. Nine questions with Eric Oliver, part of the University of Chicago Podcast Network. Mike Minnis in the first half, we talked about your research about productivity and audits and how they're connected the more audited or the better audits, the more gap audits you have, the more productive you are. The managers who get good accounts are more productive than those who don't get good accounts. How? How are they doing this? And that's what I think we need to turn to now. What is actually happening with a manager gets audited accounts or what do they do with the information that makes it better?

Mike Minnis: So there's two channels we explore in this paper. One of them is trying to get to this causal channel. If you get an audited gap set of financial statements, will you improve your productivity? Let me tell you a little bit about why that channel may manifest. So if you've got an outside accountant who comes in, normally what we think of as an audit is the task at hand is to say that manager's financial statements are materially in accordance with generally accepted accounting principles. That's the job of an audit. But anybody who's done auditing before knows that there's a lot that goes into that. One of which is auditors come in and do what they call tests of controls. In other words, what are the processes and mechanisms you put in place to generate your set of financial statements to begin with? It's not just auditing that that number is true, but how did that number get there?

How did you manage your inventory, your equipment and so forth to get those balances on those balance sheets? So when I was an auditor, I walked the factory floor. I was counting inventory. I was seeing how the machines were working. And good CPAs should be helping managers in that process to say, "Hey, you know when I noticed when we were doing the inventory checks, when we were doing the equipment checks, when we were testing your employee benefit systems, do you know that you've got problems X, Y, and Z that we found in that process?" In other words, they should be doing things to help generate both better information and identify potential processes where inventory leakage could be occurring. Inventory walking out the back door, for example, or other types of things that might hurt productivity because good processes aren't in place. They check those processes.

Hal Weitzman: So it's not just managers taking information and saying, "Oh, we can do this and that with it." It's actually finding out some of the inefficiencies from the audit process.

Mike Minnis: Correct. And I think -

Hal Weitzman: You are obviously an excellent auditor, Mike, so I know you are very diligent, but is that standard practice what you're talking about?

Mike Minnis: It absolutely should be, yes. And one of the things that we could talk about, for example, is oftentimes when we see practices that can describe why firms have higher levels of productivity than others, the immediate economist should be asking, "If it's so good, why isn't everybody doing it?" And there's a couple of pieces to that. One of which audits aren't free. They're a costly good, and managers have to decide between those costs and benefit trade-offs of doing it. My hypothesis is, I think, the costs though are very salient. When you get a bid from a CPA firm to do an audit for you, you see what that price tag is. You know it's going to take hours of time from your employees to deal with the auditors and you write them a check. One of the things I've always thought over time in both academics and in practice was the benefits are less salient.

Oftentimes managers think, okay, maybe my bank wants it for me to do it. And I've talked to some bankers before. My dad was a banker at one point in time, and he had said that one of the borrowers came and threw the audited set of financial statements on his desk saying, "I've not read this thing. You asked for it, here it is." Meaning that I think there are managers out there who don't appreciate the benefits that could come from better improved information processes. And so the costs are salient, the benefits are not so much salient. And I would say that if you are working with a CPA and they're not helping you with those processes, revisit with your CPA or revisit a different CPA to look at how that's working because it shouldn't just be a purely cost-driven exercise, but there should be benefits associated with your better information set that's going on.

So better information is one of them for the managers and they can reallocate people and reallocate across divisions because they have that information set. So that's one channel we investigate and we could talk about how we did that. A second channel for our main headline finding of more productive firms are more likely to have audited gap statements is purely a reporting channel. And that's the following sense. CPAs make you report or cause you to report more accurately what your set of financial results are. And if we step back for one second, note we are looking at the private US firm setting here, which I think is important. And what's unique about this setting and why I've done so much examination as an accounting researcher in this setting is they're not required. That's why we have variation in who gets an audited set of financial statements to begin with.

US firms are generally speaking not required to get an audited set of gap financial statements. They can choose to do so. And if they do so, then presumably their set of financial results are more accurate with respect to both input and output.

Hal Weitzman: And as you say, they might choose to do so because a lender says you have to get them.

Mike Minnis: There could be lots of reasons, absolutely. Lenders being a key reason as to why they may want to do it. But conditional on doing it, if we see one that has a good set of financial reports and one that doesn't, let's think about in the private firm setting, there may be incentives to actually under-report productivity. And you think about taxes, a tax motivated reason. So privately held companies who don't have outside shareholders and therefore maybe don't have incentives to show really good productivity results may want to show lower productivity results. In other words, lower revenue for a given set of costs to understate profits, legally or otherwise, because they want to reduce their tax burden. And so they'll bring in expenses faster and do other things that under-report productivity potentially. Presumably, an auditor would come in and help clean that up, make you really force you to report all of your productivity that you had.

So we tested this channel by noting the fact that corporations are taxed differently across states. The federal tax system is the same for all companies, but for example, California has a different corporate income tax system than Texas. In California, the incentives are very high to under-report your production because you would get taxed at a higher rate than in Texas. And so to test this channel as to whether or not these incentives to report your full productivity hold in the data, we use one of our data sets that we have state level data on to find that indeed our results are much stronger in California than they are in Texas. In other words, when auditors come in to a California firm, they really cause them to report everything appropriately. Whereas unaudited firms in California don't have that outside check from a CPA and may under-report their level of productivity.

It's kind of consistent with our results. So that's the way we kind of think about the second channel. In other words, accounting systems affect the data that empiricists use when they're trying to measure productivity. And firms have incentives to report different levels of productivity and auditors affect that reported level. So the two channels, let me step back to the first channel on that information kind of channel approach. How could they actually help them be more productive? This one's kind of hard to tease apart in the data because we don't know the counterfactual. We don't randomly assign audits. So we did a couple of things on this one. One of them is for one of our data sets, we had a panel. We could see a firm over time, very short panel of only three years. So we looked at those that newly brought in an auditor.

So we saw that they had a year without an auditor, we had an auditor and then carried forward with an audit. And what we found was if it's purely a measurement effect, in other words, all it is that auditors are coming in and affecting the reported numbers in the set of financial results, then most of the effect should just happen in that first year. You weren't audited, now you are. Now you're reporting changes and there'll be a big change in the first year and not much in the second year. The alternative hypothesis to that is a learning channel. We get a set of audited financial statements. We start to learn about our processes because they came in and tested these controls and worked with us. And a learning channel would predict changes subsequently in future years. And indeed what we find is very little initial productivity benefit to that audit in the first year we see it, and a jump in those that got audited versus those that didn't change to an audit in the subsequent years.

So in other words, some evidence that it's not just purely a reporting channel that would just come through with the way auditors cause managers to change the reports, but some evidence of a learning channel that's coming through because it's done later in the process.

Hal Weitzman: And something else that you've found is that younger companies benefit more, right?

Mike Minnis: So the younger companies, so again, if you're having a hypothesis that the channel is going to come through a learning process, try to find firms that need to learn the most.

Hal Weitzman: Correct. Right.

Mike Minnis: Those that are established and have good sets of knowledge, maybe a good CFO in place already may benefit less from it. So younger firms, which probably don't have good processes in place yet, seem to benefit more, which again is supporting a learning channel mechanism.

Hal Weitzman: Right. And to what extent is this... It strikes me the opposite is also true that you have very established companies that have just had terrible reporting for years. So to what extent is the gain that you measure in productivity to do with them actually doing something different versus just cleaning up, as you said?

Mike Minnis: Yeah, this is a good question. I think that there's lots to look at in terms of age of firm. And a lot of things that we think about, the initial thing that any question and we go back to the beginning of our discussion, the selection effect, because it's an endogenous choice, those that are going to opt into it... There's a couple of things to it. One, those that are going to opt into it are almost certainly going to have bigger benefits or realize bigger benefits to begin with. Or those that didn't opt into it should have been doing it in the first place. So the treatment effect on the treated in some sense, if you want to use a nerdy term for it, is going to be much larger than the average treatment effect of just any average firm getting this, to be clear because those firms who could most benefit from it very likely are the ones that opted into it to begin with.

Having said that, I do find it interesting in anecdotal conversations that I have with both past clients that I'd had and neighbors and so forth that are CFOs of relatively large privately held companies, and they don't get financial statement audits. And it's for a variety of reasons. And I think those reasons are very interesting and under-explored in the sense of how do those managers think about getting a financial statement audit? Why don't they do it? Some of it, quite frankly, is very interesting. They just have a distaste for outside accountants coming in and telling them what to do. It just straight up, they don't want to spend one more nickel on an accountant coming in and measuring when we should be spending more-

Hal Weitzman: Well, like you say, they don't see the value, they only see the cost.

Mike Minnis: They don't see the value, they only see the cost. Having said that, one additional dimension which we can't measure, which I'd like to measure here too, is just the quality of the CFO. Oftentimes when I did the fractional CFO, my clients didn't get audits either, but in part that was because I was there. The bank relied on and trusted the set of information I was helping them prepare. And there's probably a good substitution effect between internal accounting capabilities and external auditing. And I think that would be a great other area to explore as to where that trade-off is being made.

Hal Weitzman: I want to turn to policy because as you say, there's this massive gap in America between private company reporting and public company reporting, which is pretty burdensome. Private companies don't have to report anything. So what does your research suggest? Suggest there's a big economic effect from actually bringing in more reporting. So should we mandate? Should we encourage private companies to report more?

Mike Minnis: I would say the paper says almost nothing about policy in the following sense. I think it's just like any other management practice and companies should opt and choose how they best want to think about this choice. Good companies may choose to do it. Other companies may substitute and have good internal controls. Just because we find an effect, there's nothing that suggests there's an externality or rather some government step that should be taken here because some companies choose not to do it. There's another paper of mine that we should maybe come back in here with where we look at a mandated audit setting in the US with broker dealers. And the point of that paper is we find that actually when there's a mandate, companies who don't really want an audit go to a cheap auditor who probably does very little for them in the first place.

So I suspect if this is a setting kind of like that, if we force firms to do it, we just find a lot of companies that still never wanted to spend any money on this. And there will be auditors that are going to supply that low level of quality, which effectively isn't going to do anything in the first place. So I don't think this says much for an audit mandate in the US setting. Just because we found that there's good productivity results. I don't know that we've been solving any problem.

Hal Weitzman: Right. Okay. Well, Mike, this has been fascinating. Thank you so much for coming on the Chicago Booth Review Podcast.

Mike Minnis: Thanks for having me.

Hal Weitzman: That's it for this episode of the Chicago Booth Review Podcast, part of the University of Chicago Podcast Network. For more research, analysis and insights, visit our website, chicagobooth.edu/review. When you're there, sign up for our weekly newsletter so you never miss the latest in business-focused academic research. This episode was produced by Josh Stunkel. If you enjoyed it, please subscribe and please do leave us a five-star review. Until next time, I'm Hal Weitzman. Thanks for listening.