The NYT has an interesting report on the new wave of startups using “big data” to assess individual creditworthiness. However, the story is rather too cute by half:
By law, lenders cannot discriminate against loan applicants on the basis of race, religion, national origin, sex, marital status, age or the receipt of public assistance. Big-data lending, though, relies on software algorithms largely working on their own and learning as they go.
The danger is that with so much data and so much complexity, an automated system is in control. The software could end up discriminating against certain racial or ethnic groups without being programmed to do so.
The data scientists focus on finding reliable correlations in the data rather than trying to determine why, for instance, proper capitalization may be a hint of creditworthiness.
The problem with these “reliable correlations” isn’t that they’re wrong – it’s that they are virtually guaranteed to correlate with race, religion, national origin, sex, marital status, age, and the receipt of public assistance. While evaluation criteria like whether people capitalize names on forms are seemingly neutral, there is a reason that big credit bureaus don’t use them. The predictable consequence would be systematically higher costs of borrowing for disadvantaged groups, and the end result would look a lot like discrimination based on suspect categories. As a society, we have generally decided that access to credit should be more equitable, at the cost of being somewhat harder to forecast.
The startups featured here are trying to make an end-run around these anti-discrimination laws. Which is fine, I suppose. They are small outfits and so can embrace tactics that Equifax etc. would shun as unacceptably risky. My guess is that the DOJ will mostly ignore them as long as they remain confined to mostly niche markets, but would intervene in a big way if they become standard credit scorers. However, I think reporters should be clear on the issue and stakes involved. Discriminatory access to capital isn’t some unforeseen side-effect of better credit-scoring based on “big data”; it’s the entire point.
There’s a good reason why half the startups in the world call themselves “Uber for _______”. However, something about the phrase has been bothering me a lot, and I’ve realized it is about the different dynamics of capital and labor. Uber is about more than just car service; it is a testbed for the idea of a better capital marketplace. Instead of purchasing capital (e.g., buying a car), users rent it with their phone. Social welfare can be increased because the cost of capital is amortized across near-continual usage instead of twice a day when the car is used for commuting. The surplus is large enough that both Uber and the driver can take a cut while the user still saves money. It’s a great idea! It also works even better in many other sectors because Uber is so heavily labor-dependent and so introduces a high variable cost that induces a lower limit on available cost savings. In other applications where the user is purely renting capital (e.g., Airbnb), the clearing price might be very, very low. The user gets a benefit, the capital owner gets rents, and the platform gets a cut: social welfare is massively increased.
It works conceptually less well when the user is hiring labor – e.g., delivery, cleaning, grocery-shopping. The laborer requires pay, which means that there are variable costs. You must pay the laborer enough to make the task worth her while, because she has the opportunity cost of doing your task instead of something else. This cost structure is a huge difference from renting unused capital (e.g. a car or house), where the opportunity cost is zero or negative. There’s no reason to believe that platforms can actually provide labor-intensive services more cheaply than traditional methods. They might increase social welfare by improving matching and making the market more liquid, but that’s a marginal improvement compared to utilizing capital much more efficiently. This liquidity advantage might allow some of these startups to gain market share, but it won’t throw off huge profits because it has to be cost-competitive with existing incumbents. While some of them might be able to offer lower costs than incumbents now, that’s due to both the slack labor market and subsidies from VCs.
In short: we should expect general success of “Uber for capital”, and much slower adoption of “Uber for labor”. If you want personal service delivered to your door, you should expect to pay the same premium whether or not you book it via an app.
According to internal eBay testing, no. Randomized controlled trials with Google search engine marketing provided little to no results, and it seems like a large portion of their advertising budget is just wasted. Their budget mostly goes to show ads to those who would still have ended up at eBay anyway. As Tim Fernholz notes, there are a few gigantic caveats:
- Your mileage may vary. eBay is a gigantic retailer, and if somebody has not heard of eBay in the year 2014, they are not a potential eBay customer – they’re also probably not Googling things. This isn’t true if you’re a nobody, where effective search-engine marketing is often the single best way to spread awareness of the product.
- Long-term effects might be important. The tests show only immediate results – whereas in the long term, cutting down advertising might hurt sales a great deal. But running even very short-term RCTs at a company like eBay can be very difficult, it’s unlikely they’ll be able to do this for a year.
I would add a further point: eBay’s advertising makes competing with eBay more costly. Partially there is just the opportunity cost – your customers see ads that aren’t yours. More importantly, you’re helping your competitors’ bottom lines. Search engine ads are allocated via a competitive marketplace with competing firms bidding to place advertisements. Every bid that eBay doesn’t place makes the marketplace less competitive and allows their competitors to place ads more cheaply. This in turn makes their advertising more cost-effective. In many areas, for example online auctions, eBay is such a huge player that the spot prices for ads will probably dive a great deal, juicing their competitor’s return on investment. In the long term, that’s a serious competitive threat, which it is entirely within eBay’s power to avoid. Even if the return on advertising to them is low or negative, it might still make sense for eBay to spend billions on search engine marketing simply because they are the biggest and can best afford the cost.
Search engine marketing is an asymmetric weapon in a number of ways. It is more useful to the new and weak, and can drive growth very quickly. But when competing with an incumbent in a large category, it may well be too expensive. It also suggests that new consumer-facing startups competing with digital-native incumbents (e.g., eBay) will face systematically high marketing costs that will require massive amounts of capital. Interestingly, despite this logic and some high-profile news to the contrary, there does not appear to be a long-term upwards trend in the size of venture capital funding rounds. This might be an issue of incomparability – for example, perhaps in recent years staff-heavy enterprise startups have been supplanted by thinly-staffed consumer startups that are plugging a greater share of money into advertising. Impossible to say with the data publicly available.
If you are trying to strategically decide what kind of company to start or what market to enter, the takeaway seems clear – the idea of easily scaling up to competitive size with an established incumbent through SEM is probably an illusion. You will face systematically higher costs than you expect, and will need to deploy more of your capital than you think to advertising instead of staffing and product. As for Google, it doesn’t seem like they should be that concerned. The logic of the situation clearly suggests that even if advertising doesn’t work, the money should keep flowing in for the foreseeable future, either from established firms or heavily leveraged VC-backed startups but mostly both.
The startup world: an extremely elaborate mechanism to redistribute teachers’ retirement money to Google.
Warning – this is a post without empirical evidence.*
One of the most frustrating things ever is hearing the argument that tax incentives will bring startups to Our Depressed Rust-Belt Mid-American City. First of all, startups won’t replace the thousands of jobs lost when the old
cancer paper mill closed down – startups will employ far fewer people, and they will be mostly importing their talent from other cities/states/countries. That’s not terrible – those employees will still need meals cooked, lawns mowed, clothes cleaned, etc. The ancillary jobs of a thriving startup scene are often decent ones – but the good high-paying jobs won’t go to locals. No, the real issue with these policies is thinking that startups make decisions based on taxes.
A startup is a machine designed to turn ideas into products, and sometimes products into revenue, and products + revenue into growth. Notice a very important word that isn’t in that description? Profits. Very few startups are profitable from the get-go, and most aren’t meant to be – even the ones with massively positive cash flows show little or negative accounting profit. That’s because all of that cash is shoveled into growth. People starting startups couldn’t give two shakes about the statutory corporate tax rate because you need profits for taxes, and profits are unimaginably far away. When and if a startup becomes big enough to become profitable, it’ll already be domiciled someplace exotic with nice weather, secretive banking laws, and even more flexible tax codes.
There are tax incentives startups would care about a lot – incentives that make hiring talent cheaper. Rebates for payroll taxes are one very promising avenue. Today, payroll taxes are split half and half between employer and employee – for employers, it’s a tax on hiring people, and for employees it’s a flat tax on pay. If a local government offered payroll tax rebates to startups, that’s huge – it makes it cheaper for you to employ someone, and it makes their salary worth more. That’s actually a competitive advantage that could make it more appealing for a startup to relocate to Our Struggling Example Of Faded Mid-Century Glory And Metaphor For America’s Decline.
Tax incentives are good for existing businesses – but a narrow subset. Mid-sized businesses (small enough to feasibly relocate) with healthy profits (so the tax rate matters). Those are nice to have – but they’re hardly startups, and they’re not the type that grow quickly enough to revolutionize your city’s economy. Cutting corporate tax rates is basically orthogonal to the goal of encouraging startups, and I wish this tired old cliche were tossed out.
*: Don’t you wish more people warned you?
Techcrunch is the Valley rag – it’s always enthusiastic, talking about “disruption”, and bursting with heartwarming tales of success, entrepreneurship, and SoLoMo greatness. It is…tiring sometimes. But like any good rag, it’s essential – it’s the best source for news on product releases, rounds of funding, and the occasional high-profile scandal. And it can put out some pretty good writing from its writers and contributors. I really enjoyed a piece today by Glenn Kelman, CEO of Redfin, on the deleterious effect that success is having on the young tech community in the Bay. It’s a great example of what can be sometimes inspiring about the Valley.
The complaint is that a life of perk-laden engineering spots and eight-figure exits without a dollar of revenue has made the youth soft. Who knows if this is true – frankly, I don’t care. He’s talking about my generation, and most of the ones I know are ambitious enough for ten. Kelman seems self-aware enough; he may realize that this is always the complaint that the older generation makes about those who come after. But the value system that he speaks of is kind of touching.
One reason I love it here, and a reason that history-averse Californians never understand, is how the mindset here often seems like an oddly-askew version of Weber’s “Protestant Work Ethic“. Not in the theological aspect of it, of course. Or in the libertine attitude towards behavior. But in the elevation of personal industry and personal austerity as the ideal. For all of Facebook’s long rise, profiles of Zuckerberg inevitably and adoringly noted the disjoint between his vast fortune and his modest living circumstances. The same attitude towards personal austerity also commonly noted amongst many of the most successful founders here, and Larry Ellison’s incredibly conspicuous consumption is discussed with vague and slightly confused distaste.
It’s just a very notable difference in attitude between San Francisco and New York, America’s other great capital of capitalism. Here, the idealized capitalist is a vicious and hungry entrepreneur, building with his hands and a terminal window, clad in jeans and a flannel shirt and with vast but dematerialized riches. There, the idealized capitalist is a vicious and hungry financier, directing flows of money and profiting off his cleverness, and living the high life afforded by his vast and very material riches. Many of the older generation like Larry Page and Sergey Brin will eventually begin living the high life, but it’s never celebrated the way that conspicuous consumption is in New York City.
Of course, Californians don’t know or care for history and so most don’t see the resemblance. But for me, it’s time to go re-read some Weber.
I got locked out of my apartment this last weekend, and I must say it’s been a long long time since I’ve had a customer experience so terrible. When you are calling a locksmith, it’s generally after the shit has pretty much already hit the fan – you’re desperate, you are probably without cash, and you are totally captive. Especially if it is an inconvenient time, and isn’t it always, your ability to comparison-shop is extremely limited. And you have exactly no choice in the matter.
This is exactly the type of customer experience that “disruptors” love to talk about, but it’s not at all clear what the easiest way to do this is. It’s a problem that relies on individual skilled labor by the locksmith in situ, at any time at any location. While the second part of that is something that is a pretty solvable problem, “in situ individual physical skilled labor” is pretty much the definition of something that is terribly-suited to be solved with software or hardware. You know, the opposite of scalable. Which is why locksmithing isn’t a big business despite the extortionary 3-digit price I was charged.
Here’s my suggestion – you’re not actually paying for the service here, you’re paying for a key that will fit the lock. Unfortunately, the best way that we’ve had of creating that key up until today was in situ individual physical skilled labor. But this is exactly the type of problem that I would suggest is solvable with 3D printing – the right shape at the right time. Once we can figure out the right hardware, software and operational infrastructure to put around the technology, the locksmith is a dead profession and we’ll all be better off.
Other than locksmiths, of course – but in case you can’t tell I don’t think too highly of their line of work.
Conservatives love to harp on the (mainly imaginary) regulatory excesses of Barack Hussein Obama, which have apparently made it impossible to start a small business in this country. Which, living in San Francisco, is news to me. However, there are actually countries in which the crushing weight of government regulation make it impossible to start a successful small business, and not just North Korea. France has a thriving economy, but it has such a mass of red tape, and such inflexible labor practices, that it is actually extremely difficult to start a thriving startup in the Silicon Valley mode. Go read the Times piece – it’s an interesting look at what the imaginary “Obamanomics” would look like in terms of the consequences for entrepreneurship. Here’s my question – does this materially affect economic growth?
While this sounds crazy, I think we ought to consider France as a component of the global economy, and even just of the regional European economy. France has many large corporations and highly skilled managers and engineers, and its economy is doing just fine. It makes perfect sense if you consider that entrepreneurship, like all other economic activity, is governed by comparative advantage. As the article mentions, French culture is actually pretty hostile to entrepreneurship and aspiring entrepreneurs are likely to flee to London (and probably Berlin, startup hub of Europe). It sure seems like France’s comparative advantage lies in Areva (the gigantic state-run nuclear company) rather than letting a thousand flowers bloom. And the aspiring French entrepreneurs in turn go to the startup hubs which perform better due to the influx of French entrepreneurs.
It’s clear that having a startup hub is great for the local economy. But no one knows how to start one from scratch, which suggests it is difficult. Furthermore, the network effects thought to drive the prosperity of these hubs have positive returns to scale…which suggests that with a fixed number of entrepreneurs divided among a greater number of hubs, the total level of value creation drops. A greater number of startups in Paris would bring some incremental activity…but does it outweigh the reduced productivity of London’s startup scene? Zuckerberg moved out West because it was the only place Facebook could become Facebook – and while the money flowed to Menlo Park, the consumer surplus flowed to the whole world and everyone became better off.
Plus, there are of course good reasons for France’s restrictive labor laws – they greatly enhance the quality of life for the non-entrepreneur in France. So this article is illuminating of the choice between economic comfort and economic dynamism, but it’s not at all clear that France is making the wrong tradeoff here.