Wharton’s Lorin Hitt and doctoral student Matthew Caulfield discuss the state of U.S. voting technology.

The voting process in the United States has come into question in the last couple of decades. It started during the 2000 election of George W. Bush, when a hanging chad on a paper ballot prompted speculation about tampering. Rumors of tampering in other elections have cast doubt over the security and effectiveness of voting technology, which is why innovation in the system is more important than ever. A new report by the Penn Wharton Public Policy Initiative looks at the state of the business of voting. Wharton operations, information and decisions professor Lorin Hitt, who led the research, joined doctoral student Matthew Caulfield to discuss the report on the Knowledge at Wharton show on Wharton Business Radio on SiriusXM channel 111. (Listen to the podcast at the top of this page.)

An edited transcript of the conversation follows.

Knowledge at Wharton: What drew you to this research?

Matthew Caulfield: It was something that was part of my senior year at Wharton as an undergrad, and I found the intersection of policy and business so poignant in this particular project. You have voting technology, which is the bedrock of our democracy, and also this sort of business aspect where it’s provided by private-sector industry.

Lorin Hitt: It’s a sector where technology has the potential to create a lot of value, and that’s been the focus of my research for a very long time. Thinking about how technology can drive innovation is a core part of what we do here. I’m interested in getting involved in that where it can make a big impact.

Knowledge at Wharton: What do you think the state of the voting system is right now in the United States?

Caulfield: The current state was profiled by a recent report by the Brennan Center for Justice that said there’s a crisis in voting technology — that across the country voting technology is literally falling apart. County officials are driven to buy replacement parts from Japan online because they’re no longer provided for their current election technology. They’re trying to suck the money out of state and local governments in order to buy new technology, so it’s really in a sorry state right now.

Knowledge at Wharton: Lorin, explain the process that you and your students went through to put this research together.

“This is an unusual industry where you may have as many as 10,000 individual decision-makers.” –Lorin Hitt

Hitt: We started out just trying to compile some basic facts about the industry. One of the challenges has been that there’s not data available– or at least not data available in a consolidated way — despite the fact that a lot of organizations are interested in different pieces of the process. So, the first step was to bring together some facts about the state of the industry. Who are the competitors? Where do they participate? Do they make money? What kind of revenue flow can you expect? What’s the funding structure?

We focused on that first, and then we spent some time interviewing people and talking with other participants in the industry to get a sense for what some of the concerns were that arise from those factors and what are some of the innovative ways in which people have responded and tried to do better.

Knowledge at Wharton: What are some of the major concerns?

Hitt: Well, the concerns depend on whom you’re talking to. The folks who are voting are concerned about the fact that there are long lines, it’s operationally poor, and they’re concerned that their vote’s not being properly counted.

The election officials are just worried that they’re not going to run a smooth process, that technology is antiquated. It tends to break down at the wrong time. They’re not comfortable that they’re able to pull it off as effectively as they would like, and they’d like to move forward but don’t know quite how.

Finally, from the perspective of the industry, it’s an industry that is difficult to make money in. The folks who are trying to think about how to move the industry forward have this tradeoff [with] an industry that doesn’t have a lot of stable revenue coming in from the outside. Every time something gets replaced, you have to fight for the money. They want to invest in providing this new technology. At the same time, they’re heavily cost constrained and revenue constrained because there’s not the stable funding stream. There’s also a very large cost structure associated with innovation to get a system certified and in production. That could take years and cost millions of dollars.

Knowledge at Wharton: It appears from looking at the reporting that there are a variety of companies involved in this, but three main companies control the majority of the market right now. Is that correct?

Caulfield: Yes, that’s correct. As Lorin said, it’s very hard to get data on this market because of how small it is. Previous reports by Caltech and MIT have estimated that it’s only $300 million to $350 million per year. While that sounds like a lot of money, it’s peanuts compared to what other industries, most consulting agencies, operate in. So, we had to make estimates on revenue based on voter reach.

We tried to look at the technology that’s in different counties across the country, see how many voters they reached and then estimate revenue on that basis. We found that three companies reach over 92% of voters, with one of the companies reaching nearly half of them. It really is a concentrated market.

Knowledge at Wharton: Is there a geographic element to the companies that are involved in this?

Caulfield: Yes. We don’t analyze that much in the report, but one hypothesis is that a lot of states and counties look to neighboring states and counties to see what technology they’re using. One of the problems that we address in the report is switching costs. If you’re only buying election technology every 10 years, unless you have a veteran who’s been there a decade and a half, who’s actually bought it before, it’s a new process to you. Part of the best heuristic that people use is neighboring states.

Knowledge at Wharton: Is this an industry that could get to a point where we have one system for all 50 states?

Caulfield: We talk about the regulatory environment, and one of the issues with that idea, which a lot of people have sympathy towards, is federalist concerns. States are typically delegated these responsibilities. Now, we do have a federal commission called the Election Assistance Commission, which is under fire right now in Congress and has been for a decade, especially by conservatives who try to offer voluntary voting system guidelines that states can choose to adopt or not adopt. There are certified laboratories that people can test in, but that’s really the extent to which we can have that much of a centralized system. It’s fundamentally the states that sign on.

Hitt: This is a problem that was solved in corporate information systems decades ago, where they discovered that you could build package software systems and implementation practices, where one system could serve a very large number of companies in different ways. I think the market is potentially moving in that direction. But at the same time, you have a lot of different demands from different jurisdictions. It’s surprising how concerned people are about voting in their own particular way, and you’ve got certain jurisdictions that have multilingual voting and a lot of specialized requirements. It may take a while to get to that point, but the idea would be a limited number of very broad, customizable systems should be available. That’s the way most corporate IT works right now.

Knowledge at Wharton: Matt mentioned machines that are failing or in need of repairs. How significant of a problem is that?

Hitt: Unfortunately, it’s a fairly significant problem at this very moment, mainly because of the structure of the way funding has worked. After the hanging chad problem, there was an influx of money. About $3 billion was allocated to states, and that was spent on modernizing the systems. These systems have about a 10- to 20-year life, but the average life is about 15 years. We’re sitting right on the point where these systems are becoming obsolete. As a result, they’re going to have to be repaired or modified or maintained, and a lot of the technology that went into these systems is simply no longer available, especially now that it’s more reliant on computers.

Knowledge at Wharton: But haven’t all 50 states upgraded their systems within the last 10 to 15 years?

Caulfield: No. There’s one thing to say that some states have sort of centralized this acquisition process, such as New Mexico. It has made appropriations and bought new voting machines for all their counties. But many of these counties buy on a county-based level. So, even if you find a state where a bunch of the counties have found the money to buy new technology, many haven’t. We rely heavily on the Brennan Center report that attacks the specific issue that outlines this crisis.

Knowledge at Wharton: How prevalent are these coalitions, as you refer to them in the reporting, across the United States right now?

“Some officials may not even see that there is a crisis in election technology.” –Matthew Caulfield

Caulfield: It’s hard to say. From our experience, it would seem not prevalent. But we can’t say that definitively. It’s really just a way that counties have banded together in certain instances when they have common needs in order to deal with the current market to get features that they couldn’t otherwise get on their own.

The case we profile in the report are three counties in Florida where they need multilingual voting, and the current options simply didn’t offer that option at a cheap price. They got together and tried to form a buyer coalition, and that really worked out for them.

Knowledge at Wharton: When you’re talking about a country as diverse as the U.S., you need all kinds of different languages for people to be able to vote.

Hitt: That’s right. The larger jurisdictions face a number of these kinds of problems, especially the languages, as well as the fact that they get very large ballot lists. This combination makes their requirements very different than standard, and that’s what’s been some of the challenge. It’s finding a technology vendor willing to invest in serving their specific needs that are different than the needs of a smaller jurisdiction.

Knowledge at Wharton: What’s the state of innovation in this industry? As we move along here, new systems and upgrades are going to be needed. You’re going to need a level of innovation in this industry just like you would in a variety of industries.

Hitt: Right, so there’s this sort of regular progress of companies trying and implementing these upgrades and extensions of their existing systems. That’s been the way the industry has kind of ambled along, with small amounts of innovation over time. There are a few models that people are pursuing now. In particular, there’s a fair amount of interest in open source technology, especially on the software side, that would enable them to take advantage of a common operating system and then use off-the-shelf hardware. The hope is that we’d be able to bring costs down and make it more available. There are people interested in pushing in that direction, some coalitions trying to pursue that strategy, as well as some outside parties who are trying to make a business out of doing so.

Knowledge at Wharton: You have people in these systems and in these states that run these voting systems that have been in place for a while. They have a level of comfort and success with a particular system and don’t necessarily want to rock the boat.

Caulfield: Yes, some officials may not even see that there is a crisis in election technology. But they are also disempowered in certain ways. One is funding, which we’ve already alluded to. The other way is the current regulatory structure of the state. We’re really worried about voting machines, right? We want them to be certified to very high standards, but this makes innovation somewhat difficult. You could innovate this whole new system, apply for certification, and all of a sudden that’s not certified and you’re millions of dollars in the hole.

Also, the model certification in many states has traditionally been a monolithic system, so you get a single, self-contained system, and you certify or not certify that. These moves to having off-the-shelf technology incorporated with sort of a common software base — while it worked in many other industries, it presents a problem for the certification system. I can certify a piece of off-the-shelf technology, but once I put it together with the rest of the system there might be a whole different issue. So, structures for certification also need to accommodate this innovative spirit among many officials.

Knowledge at Wharton: What do you think is the best solution?

Hitt: I think there are a number of initiatives that could come together to solve at least pieces of the problem, maybe offer an overall solution. One is the increased interoperability standards of these technologies. Most of the technologies in use today aren’t interoperable. Folks like the Election Assistance Corporation or the National Institute of Standards and Technology (NIST) are trying to come up with interoperability standards, and that will certainly help. An open source foundation for these technologies might also help to solve some of the problems. Just finding a stable funding stream that encourages people to make these investments and spend money on innovating, and a process on the other side that gets these things certified and into production faster, would combine to increase the rate of innovation significantly.

Knowledge at Wharton: But how prevalent is that level of innovation? How frequently are people thinking about ways to adjust and tweak these systems to make them more operable across various platforms?

Hitt: I think there’s less of that from the existing industry, mainly because they’re used to working in the regulatory framework they have to work with, where innovation is somewhat penalized. But there are a number of outside firms, and it wouldn’t be surprising if they would switch over. If it was easier, for example, to get a system certified, you’d see more innovation from the existing industry. Then there’s a number of outside organizations coming in and trying to offer either an open source solution or pieces of the solution that could potentially be part of a broader system.

“If people started replacing these systems at a rapid pace, that might fundamentally change the profitability.” –Lorin Hitt

Caulfield: I recognize Lorin’s point. I think what we really do need is a paradigm shift, the entrance of whether it’s open source technology or some way to accommodate these modular technologies that you’re able to pick together. A lot of people say, “Why can’t I just vote on an iPad? What’s the problem here?” There are various problems. No. 1, the huge problem is security. Throughout this whole process, we want to ensure that elections are verifiable. This has garnered support among many election technology experts for generating a paper record. We don’t want it all on software because if there’s tampering or changing the votes, we don’t have a paper record to compare and verify whether there’s actually been a security issue with the votes. While we may criticize the certification standards, we also want them to be very high.

Knowledge at Wharton: It is interesting that as far as we’ve come, paper still is an important component to this process.

Caulfield:  Right, so you could use paper without using paper to actually count the votes. One of the biggest innovations are statistical methods that allow us to look at a state or a county, look at only a selection of the paper votes, just a small sample, and say, “We are statistically confident that the outcome of this election was correct.” They don’t need to verify the count, just the outcome of the election. In many cases, there’s enough of a disparity that counting a lot of the ballots isn’t really that important.

Knowledge at Wharton: There’s been a level of consolidation among the three main companies you reported on, which take over smaller entities to try and [gain a stronger foothold], correct?

Hitt: Yes. Whenever funding comes in or a new innovation pops up, you’ll often see a regional competitor appear, maybe getting funding or attracting interest in a particular state. As soon as the funding dries up or after they get their innovation in place, one of the larger competitors will typically acquire them to expand their geographic footprint. You can watch it happen over time.

Caulfield:  Yes, in the 2000s we had issues with a large firm becoming insolvent, being bought by the largest firm in the country, Election Systems & Software, and there was an antitrust case. So, there’s this worry from the Department of Justice and they’re forced to divest. But that firm that was divested from the largest firm just went to the second-largest firm, so we still had the same sort of distribution of these three large companies.

Knowledge at Wharton: The expectation for this industry going forward is what? Increased innovation, new systems that will come forward, and eventually we will have a more uniform system?

Caulfield: Yes. I think what will be the path forward is people recognize that a crisis has happened. There are a lot of people who are very good researchers here, who are just trying to convince people that it’s worth it to spend public money on election technology whether it’s county money, state money, federal money.

But we also recognize in the report that the solution is not just throwing money at it. We have to throw smart money at it. I think a lot of people’s intuitions are going towards the right way. “Why can’t I just use an iPad?” It’s more of an idea of, “Why can’t we just use the same nuts and bolts that my really cheap parts-seller sells. Why are we spending so much money on developing voting systems?” There are multiple avenues we can take in a secure way that lower costs and make it easier for people across the counties to replace voting machines as the best way.

Knowledge at Wharton: I would think technology also could present some issues in the fact that it’s always changing. You make that big investment in that next level of technology, but something new is going to come along and then you need to reinvest again.

Hitt: I think that’s the right insight. That’s what’s got us to this point in the first place. We started deploying technology with a limited useful life, and the combination of that and the limited funding meant that we took these technologies that should last three, five, eight years and have tried to keep them going forever.

As we move to putting in more technology, the solution is to use off-the-shelf technology, which is relatively inexpensive, has a relatively shorter life but is basically replaceable. You can replace components. You can upgrade. You can buy additional pieces that fit into a system, rather than having a system that works entirely together until it completely fails and you can’t replace the parts. The idea of moving to technology that can be slowly replaced in a modular way is one of the ways you can move forward and solve that problem.

“A county does not have that many resources or time or people or even expertise to dedicate to find the right election technology.” — Matthew Caulfield

Knowledge at Wharton: Is the regulatory issue one that really needs to be tackled at the state level?

Caulfield: The way it’s structured right now is that although it’s a state-level issue, they recruit either testing guidelines from the federal government or use testing labs that have been certified from the federal government. If there is a change in Election Assistance Commission policy, that can have ripple effects across the country in terms of the way states actually certify and test their technology.

Knowledge at Wharton: You talked about how one county will use a system and the next county will pick that up. Is that happening more at the state level as well?

Caulfield: Some of it is just search costs. A county does not have that many resources or time or people or even expertise to dedicate to find the right election technology. A state might have technologists who have either resources in this field or know people who do, who can really find the right answer for their state.

Hitt: The issue with a lot of these smaller jurisdictions is that this is not their day job. This is something that comes up every couple of years or every year where they get involved. And they’re not IT professionals. They can’t make these kinds of decisions. Some of them can, but they’re not always there from period to period. They may be there one year, and a different crowd is there the next year. Having organizations and structures that have some continuity and bring it up to a higher level where they can afford to invest is potentially helpful. States can do that or coalitions can do it where they have enough resources — where they have millions of voters who are supporting and can have an infrastructure.

Knowledge at Wharton: Is it surprising to you that there isn’t more thought put into the long term?

Hitt: I think there are a lot of people who do take the long view with this. Unfortunately, they’re not always the decision-makers who have to decide what systems are deployed on the local level. This is an unusual industry where you may have as many as 10,000 individual decision-makers, county election officials, making these decisions. They rely on these outside sources to a greater and lesser extent. There are folks who are trying to find a path forward in various kinds of ways, but ultimately the decision-makers who are deploying these systems have a fair amount of say in it, and there’s lots of them with a variety of preferences and knowledge, which is making it challenging.

Knowledge at Wharton: You mentioned the profitability of this industry, which is another one of those unknowns right now.

Caulfield: Yes, even after our year of trying to find data, we found very limited data. It’s hard to extract from vendors. They don’t want to give up their balance sheets. A lot of them either are held as subsidies of public corporations, where you can’t get their financials, or it’s just privately held by a private equity firm. We really don’t know how much money they make, but we’re pretty sure that they don’t make a lot because you don’t see entrepreneurs hopping into the voting technology industry these days. It’s not like other IT sectors where there’s money to be made or demand to be increased.

Knowledge at Wharton: Whatever level of profitability there is comes from the contract with the state or with the local election board.

Hitt: Yes. It’s totally dictated by public funding, so they’re constrained on the revenue side. You can imagine in steady state as the industry goes along, most of their revenue is replacement and repair and operations. You could see the industry growing fairly dramatically if there was an influx of funds to replace these systems, and that’s the big variable that could fundamentally change the industry. If people started replacing these systems at a rapid pace, that might fundamentally change the profitability.