By Sonali Basak
R. MARTIN CHAVEZ
Senior director and former global head of securities, Goldman Sachs
If I lie awake thinking about bad things that can happen, my concerns—and this may say more about me than anything else—are almost all about cybersecurity. Can we actually rely on, for instance, the integrity of core systems?
Here’s a scenario that is not contemplated in CCAR [the Federal Reserve’s Comprehensive Capital Analysis and Review, an annual assessment of the largest U.S. banks]: What about the Fedwire? The Fedwire is the definitive central book that says, “Who’s the beneficial owner of which Treasury security?” We rely intensely on that infrastructure. And of course, the Fed has done a very great deal to have hot backups and warm backups and cold backups. I don’t know that we’re putting enough time and energy into modeling what a disruption of banking systems—core banking systems, payment systems such as the Fedwire—what that would do to our economy.
You remember from a few years back that some hackers managed to get a hold of the Swift [Society for Worldwide Interbank Financial Telecommunication] credentials of Bangladesh Bank, the central bank of Bangladesh, and caused several tens of millions of dollars to disappear from Bangladesh Bank’s master account at the Federal Reserve Bank of New York. Some of the money was recovered, but some of it seems to have disappeared into casinos in Macau—walked out the door and was never recovered. In this case it was not a failure of the Federal Reserve. Someone managed to get access to the Swift credentials of a bank that had an account at the Federal Reserve, and they drained that bank’s master account.
The whole point of being a risk manager is not to say that something’s going to happen and to be an alarmist, it’s just to open your mind to a lot of possibilities of things that could happen—and then get yourself comfortable that they’re really unlikely. So I just don’t know about all of the security arrangements of the Federal Reserve, for instance, or of the ECB [European Central Bank]. I have confidence that they take these matters with extreme seriousness, but I personally don’t know exactly what they are doing, what kinds of technologies they’re using, to safeguard the system.
Almost all U.S. Treasuries do not exist in the form of a paper certificate—over 99.9% of them exist purely in electronic form. And Treasuries are the beating heart of the global financial system. Every country has its inventory of U.S. Treasuries. Treasuries are used as collateral for everything.
And yet, if you ponder that they exist entirely in electronic form, you’ve got to really start worrying about that electronic form. I am actually less worried that [these systems] could be hacked and simply halted. The thing that I think worries me more is, could it systematically be corrupted by a hacker? So instead of having confidence in who is the beneficial owner of every Treasury, [you might wonder] in whose possession is that Treasury at every moment in time? Because that’s the core of the financial system: moving Treasuries around. But when you think about it, the Treasuries are electronic—they’re not actually being moved in physical space; there’s just a computer somewhere that’s keeping track of who owns them. And if someone could get into that record and cause us to lose confidence in who owns the Treasuries, that would be, I mean—it’s so hard to even think about that outcome—it would be so extreme and so dire.
However, I worry more about nonfinancial companies than I do about financial companies. If you looked at the pandemic, there was very little concern about the integrity and stability of banks. Think of how startling that is, right? Compare that to the financial crisis, which was all about concerns about participants in the financial ecosystem. In the current crisis, the concern has been about everybody except banks, and I would say an important reason for that is CCAR.
Should there be a CCAR equivalent for systemically important nonbanks? As we discovered in the pandemic, there’s a lot of systemically important companies. It suddenly became obvious to everybody. Without Amazon or Google or our internet service provider, our problems would become even greater. And so, do we want to have some kind of framework so that we can have confidence in nonfinancial companies in a crisis?
There’s been a lot of concern over the past few years about artificial intelligence. Will some AI take over, and then we become the AI’s pets? Well, I’m actually more worried about something that I think has already happened.
We already have massive AIs in the form of these tech companies. Their data centers are running software on those millions of computers, and collectively they are artificial intelligences. And they’re artificial intelligences that are systematically exploiting weaknesses in human psychology: our tribalism, our gullibility, or wanting to be told what to believe, our wanting to be liked, our wanting to be told that we’re right. And they’re exploiting it to the end of maximizing advertising revenue. So, yes, I think we have to ask ourselves: Should we even allow this model of targeted digital ads? I worry a lot less about subscription services. If I’m paying someone a subscription like Netflix, their job is to keep me happy so I keep paying that subscription. But you know that old saying: “If you are not paying for some product or service, then you are the product or service.” So, yeah, I spend a lot of nights lying awake thinking about the extent to which we have become the product of artificial intelligences that are selling our attention and our behavior to advertisers. I think that core business model is extremely problematic in a way that untrammeled, undercapitalized trading and inventory of risk was a problem that was part of the financial crisis.
Former co-CEO Bridgewater Associates
I think the next pandemic that’s coming is the displacement of the workforce [that’s] not being trained to participate in the economy. Dealing with that will take a lot more than a vaccine. The unskilled worker is the next pandemic.
Let me just step back. First of all, I think that it’s pretty clear if you have high degrees of unemployment, there are basic economic impacts that are not good for the nation. It puts an incredible strain on the economy. It usually only gets addressed through higher taxes on a smaller and smaller tax base. It increases the gaps between the haves and have-nots, which historically has caused more social unrest if it goes on for a period of time. We saw high levels of unemployment like in the Depression create hopelessness and despair in individuals, in families, and their communities.
I think [unemployment] happens as quickly as automation displaces people from work without concurrently retraining them and retooling them for other types of work. Does that make sense?
Look at the amount of money the U.S. spends on education per capita, and look at where it places in terms of education among developed countries. There’s a mismatch there. Throughout history, education has proven to be the vaccine for poverty and for the gaps between the haves and have-nots.
What I’m seeing is a growing gap between the haves and have-nots. The rich are getting richer, the poor are getting poorer, and we’re losing the middle class. To the extent that you have rising unemployment, you have people who can buy fewer goods and services, and you also have a need for the government to increase the tax base, including corporate taxes. The better way is for business, government, and educators to get together and say, a) “Do we all agree there’s a looming problem ahead of us?” and b) “What are the things that we can jointly do to start working on that problem?”
I’ve been talking about the issue for the last 10 years. McKinsey did a 2013 study and found that up to $9 trillion in global wage cost can be eliminated by automating a wide range of knowledge-intensive tasks such as analyzing customer credit ratings and providing financial advice. That’s 2013. I don’t know what the numbers would be today. There was also a study by the Oxford Martin program on technology and employment: Only 0.5% of the U.S. workforce is employed in industries [associated with new technologies] today that didn’t exist at the turn of the century. And it’s not just that these people aren’t a significant part of these new industries—it might be the nature of the industry—but we need to find ways of getting people more involved and better trained. And I don’t mean just college training. It might be training in different types of work such as working on rebuilding infrastructure. I think it’s a function of government, education, and business getting together.
In other words, this problem isn’t going to be solved by a vaccine. It’s not going to be solved by more military force or more security. It’s not going to be solved by those kind of things that happen more quickly. It’s going to be solved by retraining or retooling.
Let me give you an example. Companies may have to put up capital for future litigation costs. They might have to put up capital for certain regulatory things, so should regulators give companies that retrain their people a break vs. companies that don’t retrain and retool?
There’s been a lot of college graduates. And a lot of them didn’t find jobs. Did we produce more college graduates than the economy needed? But I also know that we don’t have enough electricians; we don’t have enough plumbers. Are there other skills beyond a four-year college diploma or an MBA that could be alternative ways of training or developing people for the future economy?
I don’t think companies will keep unproductive people around for very long. Here’s what I think is hopeless: If you tell someone this is your life for the next 40 years, and you’ll be doing the same thing over and over again, when you take away from people the hope to realize their potential, that is the epitome of despair.
Is this the best we can do societally? Is this the world we’re going to leave for our children?
Co-founder and co-chairman Two Sigma Investments
One thing that I worry about, which may not necessarily be something that people talk about all the time—remember, I’m a tech guy—is that we may be building a world that is not particularly designed for humans.
Henry Ford talked about it long ago. He wanted to produce cars, but he wanted to make sure that the people producing the cars could earn enough money that they could buy the cars. And you get this wonderful effect where the growth in economic output produces good jobs, which then will help grow the economic output. It becomes a virtuous cycle.
What we’re doing today is finding more and more ways to essentially reduce the need to have humans involved with work. So much of the investment in business in America is to essentially automate away human labor or, even more curiously, to devalue human labor.
The nature of work is changing. For some people it’s getting a lot better. But for probably the majority of people, it’s not really getting better. If you can’t fix this problem, what could occur is that we end up with this sort of barbell economy where a lot of people aren’t really doing all that well. Not only is that just not a good thing, but also they’ll have less money to spend to make the economy get bigger, so the Henry Ford thing won’t occur. I think this is being overlooked. And it’s happening kind of slowly.
I am not at all concerned that there’ll be a shortage of work. There will be plenty of things for people to do. The problem is, they may be things that we don’t want to pay much money for.
Are we truly building a better world for ourselves? Or are we trying to essentially optimize some number like GDP, which in the end doesn’t have all that much to do with whether or not we’re happy and fulfilled and it’s a good stable society? So we’re having more and more focus just purely on automating stuff away, trying to make things more efficient, that may not actually make your life and my life and everyone else’s lives feel better. It may not be like a human-centered society.
I don’t think people are reflecting enough, in America anyway, as to whether or not this is precisely what we want. It’s different in other countries. In Japan, for example, I think people have put more emphasis on upgrading experiences because they understand that when you go out—walking down the street, going into stores, and all that—if you make it appealing, people not only will want to go to your store, but it’s an experience, you enjoy it. So what will the modern experiences be? Are we going to all be sitting at home, wearing VR goggles, and clicking away all the time?
Social networks like Facebook, people talk about all the problems, but you know one thing that it does is it makes socializing very efficient. So you can have your 100 friends or 500 friends, and you don’t have to waste time calling people up telling them what happened. You go click, click, click, and then everyone knows about your life. It’s the most efficient way to socialize, right? Something doesn’t sound right. You know that. Do you really want these levels of efficiency?
I’m not nearly the first one to say this, but at least at a very high level, GDP is the thing everyone looks at. And, you know, quite frankly, I’m not sure that that matters all that much. Because it’s literally such a crude number. Do you really care about GDP? Is that how you’re gauging the quality of your life? No? OK. We don’t spend very much time thinking about the quality of life more broadly. You could in some sense create an ever-growing GDP but have most people saying this is pretty bad. We could end up with a lot of unhappy people. Maybe we’re there already.
A lot of people talk about, is AI dangerous? Some people have been out there saying it’s more dangerous than nuclear weapons. I think that that’s very silly. AI is a very useful tool. But the question is, how do you use it?