00:00:02
Digital innovation is giving rise to new
00:00:04
business models. Uber and Airbnb are
00:00:07
household names today when not so long
00:00:10
ago we were all learning about the
00:00:12
sharing economy. But regulations don't
00:00:15
always evolve as quickly as uh technolog
00:00:18
technological change at least that's a
00:00:21
perception. Uh so what should
00:00:23
policymakers and regulators do? Here to
00:00:27
talk about some his some of his insights
00:00:29
is uh Wharton professor Kevin Warbach.
00:00:32
He recently wrote a policy brief for the
00:00:34
pen Wharton policy public policy
00:00:37
initiative. Uh welcome Kevin. Thanks.
00:00:40
Glad to be here. So in your article you
00:00:42
mentioned something called the internet
00:00:44
of the world. Can you tell us what that
00:00:46
is? Well I think there's something big
00:00:49
going on and it's it's a bigger trend
00:00:51
than most people realize. There are
00:00:54
three trends that each in of themselves
00:00:56
is significant. One is uh what we often
00:00:59
call the sharing economy. I think it's
00:01:00
really more the ondemand economy. It's
00:01:02
not just about sharing resources, but
00:01:04
services like you mentioned like Uber
00:01:05
and Airbnb which give ondemand access to
00:01:07
resources. The second piece is the
00:01:09
internet of things, all kinds of
00:01:11
devices, billions and billions of
00:01:13
devices gettingororked. And the third is
00:01:15
big data and analytics, the ability to
00:01:18
understand and manipulate trends coming
00:01:20
out of all those devices. What those
00:01:22
three things together mean is that all
00:01:26
of the world potentially isworked. It's
00:01:28
not just that you go somewhere to a
00:01:29
computer, you go to your phone to get
00:01:31
access to information. It's that
00:01:33
potentially everything is a generator of
00:01:35
data and all that data can be integrated
00:01:37
and analyzed and processed and
00:01:39
manipulated. Um, and so what that means
00:01:41
is the kinds of trends and the kinds of
00:01:44
developments that we saw online are now
00:01:47
happening offline. they're happening to
00:01:49
things and physical objects in the world
00:01:50
as well. You also point out that the
00:01:54
scale of ondemand services is
00:01:57
potentially much greater than the legacy
00:02:00
industries they challenge. How so? So
00:02:03
there's this uh kind of cheap talk about
00:02:07
new technologies disrupting old
00:02:08
technologies. And actually the theory of
00:02:10
disruptive innovation which goes back to
00:02:11
Clayton Christensen at Harvard Business
00:02:13
School is a a serious academic theory.
00:02:15
Uh but far too often people in business
00:02:19
and entrepreneurship and the media kind
00:02:21
of use the word disruption as just kind
00:02:23
of a synonym for new technology. And the
00:02:26
reality is it's not that you have one
00:02:28
market and suddenly a bunch of new
00:02:29
companies come in and replace that
00:02:31
market. Often what happens and this is
00:02:33
what we're seeing with things like
00:02:34
ondemand economy is that the new markets
00:02:36
are different. So it's not that Uber
00:02:40
takes the taxi market and every taxi
00:02:42
gets replaced by an Uber driver. Uh in
00:02:45
fact Uber has put out some numbers for
00:02:46
the past several years that show that
00:02:48
the scale of the market they are tapping
00:02:50
into is actually much bigger. So they
00:02:53
generate far more revenue in the cities
00:02:55
where they are mature than the taxi
00:02:58
industry as a whole does. So what that
00:03:00
means is it's not just a competitive
00:03:02
threat and certainly it is a competitive
00:03:03
threat to the incumbent industries but
00:03:05
it's creating something new. It's
00:03:07
unlocking latent demand that the
00:03:10
previous approaches didn't reach.
00:03:12
You also say that um you also pointed
00:03:15
out that throughout the different techn
00:03:17
technological waves since the 1990s we
00:03:20
went through e-commerce, social media,
00:03:22
now mobile that regulations have always
00:03:25
been seen as an enemy of innovation. But
00:03:28
you say that this digital dichotomy is
00:03:31
actually misunderstood. Can you explain
00:03:32
that? Sure. So there's two pieces to it.
00:03:34
One is the term that you reference that
00:03:36
I use in the paper called the digital
00:03:38
dichconomy. That is a misunderstanding
00:03:40
that the online world is inherently
00:03:43
different from the offline world. Uh and
00:03:45
the reason that's not true is what I
00:03:46
said at the beginning. Um increasingly
00:03:49
there is no difference even if you're
00:03:50
using a physical thing. So take the Uber
00:03:52
example and it's such a perfect example.
00:03:54
Uh the per there's a person there's a
00:03:56
physical person driving a physical car
00:03:58
but from your standpoint running the app
00:04:00
and pushing a button and saying make a
00:04:02
car appear. It's as though that's
00:04:04
something that's in cyerspace. It's as
00:04:06
though it's something digital. an
00:04:08
extension of the software infrastructure
00:04:11
of Uber. Uh even though it's a physical
00:04:13
thing, a physical person driving a
00:04:15
physical car. Um and so we tend to
00:04:17
assume that there's one set of rules for
00:04:19
the real world, there's one set of rules
00:04:21
for the digital world. And that's a
00:04:23
mistake because increasingly there is
00:04:25
just the world. Software, technology,
00:04:28
networks, all these trends in what I
00:04:29
call the internet of the world are
00:04:30
affecting everything. So that's the
00:04:32
first piece. the assumption that uh we
00:04:36
can just ignore the rules of the
00:04:37
physical world because we need totally
00:04:39
new rules for the digital world. Um the
00:04:41
larger issue though uh is this question
00:04:44
of innovation and regulation and again
00:04:47
there's this common assumption that
00:04:49
innovation needs to thrive with no
00:04:52
regulation and anytime government gets
00:04:54
involved that's a check and a drain and
00:04:55
a block on innovation and that's not
00:04:57
really the case and what I talk about in
00:04:59
the article you referenced uh and the
00:05:02
larger larvy article it's based on is
00:05:04
that if you go and look at the history
00:05:06
uh how the internet developed how
00:05:08
electronic commerce developed in the
00:05:09
1990s
00:05:11
A surprising amount of the time it was
00:05:13
government action actually facilitating
00:05:15
innovation and uh the emerging startups
00:05:19
actually pushing for that government
00:05:21
innovation to help create a more
00:05:23
innovative marketplace. That's an
00:05:25
interesting point and uh in your article
00:05:28
you also pointed to uh one challenge for
00:05:31
regulators and that is um a lot of these
00:05:34
new startups don't really fit neatly
00:05:36
into industry categories. So uh and the
00:05:40
example you pull up you use is uh Uber
00:05:43
versus Skype. Can you go through that
00:05:45
example? Sure. So I should be clear.
00:05:47
It's not that regulators always get it
00:05:49
right. They make mistakes and they have
00:05:51
lots of flaws and lots of reasons why
00:05:53
they act in a certain way. Uh and we
00:05:56
should definitely criticize bad
00:05:58
regulations. Um we just shouldn't assume
00:06:01
necessarily that they are bad and
00:06:03
necessarily what what startups do is
00:06:05
good. Um the Skype and Uber comparison
00:06:07
is basically that both of them were
00:06:10
companies that when they started were
00:06:13
illegal in most jurisdictions. So Skype
00:06:15
which is the very popular internet
00:06:18
communication service originally voice
00:06:19
calling now also video and messaging and
00:06:21
so forth eventually now owned by
00:06:24
Microsoft. Uh Skype was illegal in most
00:06:26
of the world when it launched because
00:06:27
there were rules saying you could not do
00:06:30
a communication service, a telephone
00:06:33
service outside of the existing
00:06:35
regulatory infrastructure in the US
00:06:37
because of what we did. I was at the
00:06:39
Federal Communications Commission in the
00:06:40
1990s when we had to think about
00:06:42
voiceover IP. We very deliberately left
00:06:44
open the door. Even though things like
00:06:47
Skype were outside of the regulatory
00:06:49
structure, we made a conscious decision
00:06:51
to allow them to develop. And that's an
00:06:53
example where regulators consciously
00:06:56
deciding not to impose a whole set of
00:06:57
rules early on when these were nent
00:06:59
technologies allow them to grow. Um so
00:07:02
Uber is similar. Uber is illegal in most
00:07:04
of the cities where it operates. Uh and
00:07:08
the story of Skype I think is a hopeful
00:07:09
story because what happened with Skype
00:07:11
is that first of all you had regulators
00:07:13
like the FCC in the US who understood
00:07:16
that these new internet calling
00:07:18
technologies were an opportunity. They
00:07:20
were a way to lower prices and create
00:07:22
better service and new service and
00:07:23
innovation. Uh and so that we shouldn't
00:07:25
rush to impose all the traditional rules
00:07:27
on them. And then as these companies
00:07:29
grew, they were able to work with
00:07:31
regulators to address the rules that
00:07:32
were necessary. So for example, it's a
00:07:35
concern if you're using Skype or some
00:07:36
other voiceover IP service. What if you
00:07:39
want to call 911? What if you have an
00:07:40
emergency? Um those services were all
00:07:43
outside of the infrastructure of the
00:07:45
emergency calling system. And that's a
00:07:47
problem if you have your phone and which
00:07:49
app you're using determines whether if
00:07:51
you have an emergency you get through to
00:07:53
the police or the fire department. Um
00:07:55
that's a technical problem that was
00:07:57
overcome through regulators working
00:07:59
together with the companies. Um and it
00:08:01
could only be done by a willingness to
00:08:04
not immediately impose rules that would
00:08:06
have shut these companies down on day
00:08:07
one, but work through how they can
00:08:09
actually help in addressing some of
00:08:11
these concerns. And I think that's the
00:08:13
path forward for companies like Uber and
00:08:14
Airbnb as well. What we've started to
00:08:16
see is they began with this very strong
00:08:18
keep no regulation, keep it away from
00:08:20
us. Uh, and gradually we're starting to
00:08:23
see them recognize that they need to
00:08:25
work together with the regulators. But
00:08:27
don't you think there's one critical
00:08:29
difference between Uber and Skype and
00:08:31
that is, you know, with Uber there's the
00:08:33
issue of safety of the users or the
00:08:36
writers, but that's not the case with
00:08:38
Skype. So the example I gave about the
00:08:41
911 service, the emergency, that's a
00:08:42
safety issue. Um but no question uh all
00:08:45
these companies are different. Uh and
00:08:47
it's not the case that there's one set
00:08:49
of rules, there's one regulator that
00:08:51
applies to everything. Um I'm making the
00:08:53
general point that these new services
00:08:56
that spring up outside of the
00:08:58
traditional regulatory structure. Um
00:09:02
they raise public policy issues uh and
00:09:04
you can't get away from that. So Uber
00:09:06
raises a whole host of issues. It raises
00:09:08
issues about worker treatment. It raises
00:09:10
issues about safety of the people in the
00:09:12
cars as you mentioned. raises issues
00:09:13
about the Americans with Disabilities
00:09:15
Act. Do they have to provide access for
00:09:17
people in wheelchairs and so forth? All
00:09:18
these issues. Um, and the way to address
00:09:21
those issues is not to say it's new, so
00:09:23
therefore regulators stay away, they'll
00:09:25
figure it out. The way to address them
00:09:27
is to say, okay, let's look at the
00:09:28
issues. So, um, is there a way to
00:09:31
address what we really care about in
00:09:33
safety? What we care about is you don't
00:09:35
want people to get into a car and
00:09:37
someone be an axe murderer who's driving
00:09:39
the car. All right. there's different
00:09:41
ways that we can solve that problem and
00:09:43
and maybe there are solutions and in
00:09:45
fact Uber because they have all this
00:09:46
data and they've got an app and they
00:09:49
know everything about the driver maybe
00:09:50
there's a better way they can solve that
00:09:52
problem. Um so the way to work that
00:09:54
through is to have that discussion and
00:09:56
say identify what it is the regulators
00:09:58
are trying to do and figure out how we
00:10:00
can do that. The problem is initially so
00:10:03
much of the reaction was the regulators
00:10:06
are necessarily bad. um they have
00:10:08
nothing to do but to stop the
00:10:09
innovation. Uh and I think that's not a
00:10:11
helpful conversation. Uh your point is
00:10:14
that government can actually be a
00:10:16
positive force in innovative markets.
00:10:19
Can you give us more examples of that?
00:10:21
Sure. So we saw a lot of examples with
00:10:23
the growth of the internet and
00:10:25
electronic commerce starting 20 years
00:10:27
ago. Um one of them was the antitrust
00:10:30
case against Microsoft. So Microsoft was
00:10:31
the dominant company on the personal
00:10:34
computer with the operating system
00:10:35
market and lots of startup companies,
00:10:39
companies like Netscape realized if they
00:10:41
wanted to innovate, they wanted to build
00:10:43
the internet economy as we know it
00:10:44
today, you couldn't have Microsoft
00:10:46
standing there using its power at the
00:10:49
time. It's hard to realize today with
00:10:51
what's happened, the growth of Apple and
00:10:54
the growth of smartphones and so forth,
00:10:56
just how much power Microsoft had as a
00:10:58
bottleneck. Microsoft controlled access
00:11:00
to the PC and the PC was the only game
00:11:02
in town. Um, had it not been for that
00:11:06
action by the government in filing that
00:11:08
antitrust case, Microsoft may have been
00:11:10
able to warp or slow down the growth of
00:11:13
the open internet economy. And it turned
00:11:14
out most of the startups were on the
00:11:16
side of the government in that case
00:11:18
wanting to open it up. Um, a similar
00:11:20
more recent case is the fight over
00:11:22
network neutrality rules where lots and
00:11:24
lots of startup companies went to the
00:11:27
Federal Communications Commission and
00:11:29
said, "We don't want broadband
00:11:31
providers, the access providers, the
00:11:33
internet service providers or ISPs as
00:11:34
they're called to stop us from getting
00:11:37
into the market or to basically tax us
00:11:40
to say, well, you can only get to
00:11:41
customers if you pay us this special
00:11:43
fee." Um and so they were actually
00:11:46
urging government to act in order to
00:11:48
create a more open market. Great. Um you
00:11:52
also say that ondemand services would
00:11:54
bring what you call algorithmic
00:11:56
competition policy questions to the
00:11:58
four. Why is this important? What do you
00:12:00
mean by that? This is a really
00:12:02
interesting issue. So competition
00:12:04
policy, I gave the example of Microsoft
00:12:07
um is tremendously important to the
00:12:10
digital economy and uh the Microsoft
00:12:12
case was an example where there was a
00:12:15
new kind of uh business model. Microsoft
00:12:18
was one of the first to build this
00:12:21
platform network-based business model
00:12:23
where Windows benefited from all the
00:12:26
applications on top of Windows. uh but
00:12:28
Windows would always want to uh ensure
00:12:31
that none of those applications would
00:12:33
then compete with it. Um and there was
00:12:36
tremendous benefits of that model. You
00:12:37
know, Microsoft did great things for
00:12:39
innovation. Uh but uh the Microsoft case
00:12:42
put a spotlight on some of the dangers
00:12:44
and the downsides. What we're seeing now
00:12:46
with these next generation platforms,
00:12:48
these ondemand platforms is a new twist
00:12:50
on that model. Um companies like Uber
00:12:52
and Airbnb are built on algorithms. are
00:12:55
built on software that understands
00:12:56
supply and demand uh and matches people
00:12:59
on both sides of the network. Um and
00:13:02
again that's a tremendous boon for
00:13:03
competition and innovation. I'm not
00:13:05
saying it's bad by any means but it does
00:13:07
put the platform owner in a position of
00:13:09
unimaginable control. Um how do you know
00:13:13
that what you are paying for that Uber
00:13:15
ride is the efficient price? Uber says
00:13:18
well by definition it's what the
00:13:19
algorithm gives you. Well but who
00:13:20
controls the algorithm? Um, and what
00:13:23
stops the algorithm from colluding with
00:13:26
someone else's algorithm uh behind the
00:13:28
scenes to fix prices? And again, we have
00:13:30
antitrust doctrines about things like
00:13:33
price fixing, but those are based on
00:13:34
people in a smokefield room saying,
00:13:36
"Okay, you're going to charge this and
00:13:37
I'm going to charge that and if you
00:13:38
defect." Um, now it's all happening
00:13:40
silently through software. Um, and so I
00:13:43
think this is one of the great
00:13:44
competition policy challenges of our age
00:13:47
is how to prevent those kinds of
00:13:51
mechanisms from uh raising costs and
00:13:54
raising prices and hurting consumers
00:13:56
while still allowing flexibility for
00:13:57
companies to innovate and do things that
00:13:59
most of the time actually wind up uh
00:14:01
helping consumers.
00:14:03
So that brings us to the point you made
00:14:04
about algorithmic cartels. So how could
00:14:08
those come about? Uh well so again uh
00:14:11
the algorithms can talk to other
00:14:12
algorithms and we see this um already
00:14:15
you look at uh pricing on Amazon.com. So
00:14:18
Amazon has this platform that allows
00:14:20
anyone else to sell on Amazon.com and
00:14:23
they can set their price and a lot of
00:14:24
the companies that are sophisticated
00:14:25
that do this uh set their prices
00:14:28
algorithmically. So they might say
00:14:30
Amazon is charging this price
00:14:32
automatically charge 2% less than
00:14:34
Amazon's price. So when it comes up
00:14:36
they're the cheapest price. Um, but what
00:14:38
happens is you get this increasingly
00:14:40
complex uh war between the algorithms
00:14:43
because they're all basing their prices
00:14:44
on each other and so forth. And what can
00:14:47
potentially happen is uh companies
00:14:50
decide well no let's both agree we'll
00:14:52
set a price higher as opposed to
00:14:54
competing and a race to the bottom and
00:14:56
we'll both be better off but who's worse
00:14:58
off is consumers. So um that's a concern
00:15:00
that we're starting to see on platforms
00:15:02
like Amazon and it's more of a concern
00:15:05
on these digital ondemand platforms
00:15:08
where again everything is in software uh
00:15:10
and where you have lots of different
00:15:12
actors coming together um and uh we
00:15:16
don't even know what the mechanism is to
00:15:19
get access to the data to see if that's
00:15:21
what's happening. So how do you regulate
00:15:23
that? Uh well so it's a good question.
00:15:26
The first is you start to have a
00:15:27
conversation where the regulators say
00:15:29
here's what we're trying to achieve and
00:15:31
the companies say here's what we're
00:15:33
doing and you figure out what's
00:15:35
possible. Um ultimately as I said there
00:15:38
needs to be access to the data. Uh and
00:15:40
this is a great opportunity because
00:15:42
these new platforms generate tremendous
00:15:45
amounts of data. uh and if they uh and
00:15:48
they use the data internally to be more
00:15:50
efficient and to provide better service,
00:15:53
but if they could provide more
00:15:54
transparency of that data um that would
00:15:57
give regulators the opportunity to
00:16:00
identify what the market performance is
00:16:01
and this can be done in a secure way in
00:16:04
a way that doesn't uh harm them with
00:16:06
competitors and so forth. um but it's
00:16:09
actually making the regulation itself
00:16:11
more algorithm making the regulation
00:16:12
itself more datadriven which is a
00:16:14
healthy and a good thing and so I think
00:16:16
this is potentially the new model we're
00:16:19
going to come to uh but it takes the
00:16:22
company's willingness to work together
00:16:24
and not to you know make these sort of
00:16:26
great you know statements uh that oh we
00:16:28
don't need any regulation
00:16:30
so you mentioned alternatives to direct
00:16:33
direct regulation which are
00:16:35
self-regulation and what you call
00:16:37
co-regul regulation and delegated
00:16:39
regulation. Can you explain the
00:16:41
differences among all those? Yeah, these
00:16:43
are models that actually are used much
00:16:44
more widely elsewhere in the world,
00:16:46
especially in Europe for things like
00:16:48
internet content. Um, and there's a
00:16:50
whole variety of different models, but
00:16:51
basically u they start with the notion
00:16:54
that um companies individually and
00:16:57
industry collectives and industry groups
00:17:00
um potentially know the most about their
00:17:03
market. uh and if they're well-meaning,
00:17:05
they can come up with mechanisms that
00:17:07
achieve the goals of regulators without
00:17:10
government having to be intrusive and
00:17:11
without government having to be
00:17:13
inefficient. Um because regulatory
00:17:15
agencies don't have the data and they're
00:17:16
not set up to operate in that way. The
00:17:18
problem is you need some accountability.
00:17:20
So just saying let companies regulate
00:17:23
themselves is meaningless because
00:17:25
there's already always incentives for
00:17:26
companies to cheat or to game the system
00:17:29
or to basically help themselves at the
00:17:32
expense of the public. Um but there's a
00:17:34
variety of mechanisms where for example
00:17:36
government sets goals and then gives
00:17:38
industry or either individually or
00:17:40
industry groups uh uh opportunities to
00:17:42
meet it and then to report on how
00:17:44
they're doing and again to provide
00:17:45
transparency of the data. Um they're
00:17:47
mechanisms that basically say all right
00:17:50
in the first instance you have this
00:17:52
opportunity to act but if you don't act
00:17:54
in a way that we find appropriate then
00:17:57
we're going to intervene. Um and uh
00:18:00
again there's a variety of different
00:18:01
variations on these mechanisms. Uh but
00:18:04
it's an approach that says instead of
00:18:07
everything starts with the regulator the
00:18:08
regulator says yes or no before anything
00:18:11
happens in the marketplace. It says,
00:18:12
"All right, companies can come into the
00:18:14
marketplace, especially new companies."
00:18:15
Goes back to what I was saying before,
00:18:17
nent small innovators, should have lots
00:18:20
of running room. Uh because even a good
00:18:22
rule will kill them off when they're too
00:18:24
small. Um but that doesn't mean that
00:18:26
when you've got an Uber, which is a $65
00:18:29
billion company, it's one of the, you
00:18:31
know, the largest companies in the
00:18:32
world, even though it's not even public,
00:18:33
and has hundreds of millions of uh
00:18:36
people that it's serving around the
00:18:37
world. They're not a small company. um
00:18:40
they're fairly new, but but they're now
00:18:42
a large player that's having big
00:18:43
impacts. It doesn't make sense to say
00:18:45
that that approach makes sense for them
00:18:47
anymore. Um so the question is, can we
00:18:50
allow them in and then have a process to
00:18:53
then collaboratively have them work with
00:18:56
government uh and have them be more
00:18:58
transparent? And again, there's a
00:19:00
variety of ways to do that. And I'm not
00:19:02
arguing for any particular one as being
00:19:03
perfect. I'm arguing for an openness and
00:19:06
a recognition um that um regulation
00:19:10
isn't a dirty word. Any final thoughts
00:19:13
for policy makers and regulators?
00:19:17
Um so regulators have to take action
00:19:19
here too. It's not that they need to
00:19:21
just stay where they are and expect the
00:19:23
companies to come to them. Um often
00:19:25
there's lots of legacy in regulation and
00:19:27
some of it is regulators fault and some
00:19:28
of it is the fault of for example the
00:19:30
legislators that set up the rules. Um, a
00:19:32
lot of what we're seeing in these
00:19:34
markets is the need for legislative
00:19:36
change, for um, governments to change
00:19:39
the structure of the rules because the
00:19:40
rules use terms that no longer make
00:19:42
sense or they have categories that no
00:19:43
longer make sense. Um, and so there
00:19:46
needs to be a lot of dialogue between
00:19:48
industry and regulators and legislators
00:19:50
to say, all right, where are these
00:19:53
glitches? Let's fix them. Um, and
00:19:56
regulators need to be part of that and
00:19:58
not to just assume that the status quo
00:20:00
is the right approach. Um, regulators
00:20:02
also need to be open. They need to go to
00:20:03
these uh companies and say to them,
00:20:06
"Okay, we have shared goals here. We're
00:20:09
not here to put you out of business. Um,
00:20:11
but we care about consumers and we trust
00:20:13
that you do too. So, let's let's come up
00:20:14
with a solution." Um, and so it really
00:20:17
has to go both ways and and ultimately
00:20:19
this is about trust. There needs to be a
00:20:21
mutual process of generating trust
00:20:24
between these industries and the
00:20:26
regulators. Um, and in a lot of cases
00:20:29
that's lacking, but I'm hopeful and I
00:20:31
think the examples that we saw with the
00:20:32
growth of the internet really are a
00:20:33
story about um, good work on both sides
00:20:37
uh, that facilitated this extraordinary
00:20:39
explosion of innovation and wealth
00:20:41
creation that we saw. That's very well
00:20:43
said and thank you so much for joining
00:20:45
us. Thanks. My pleasure.
00:20:58
[Music]
