Friday, February 28, 2003

Heading East

Just accepted a job with Big Blue. Come the summer, I'm heading back to Boston :)

Evil Google

I've been tangentially watching Dave Winer (of Userland) and Evan Williams (of Blogger) going round and round on what Google's acquisition of Blogger means. (Full disclosure: I use Google and love it, I use BloggerPro and like it, I tried UserLand and did not like it).

Here's something I recently learned in school: venture capitalists hate to go up against Mom-and-Pop businesses. They hate it because Mom-and-Pop businesses are tough as nails, and will refuse to exit a market long after purely financial companies have pulled the plug. That's not to say venture backed companies won't swing at Mom-and-Pop dominated business sectors (eg. Staples) but it makes them nervous and afraid.

The reason Mom-and-Pop businesses are so tenacious has to do with the standard operating decision all firms have to make. If you are making an economic profit by selling above marginal cost and covering your fixed costs, you stay in business. If your price drops and you are only just barely covering your marginal cost and aren't covering your fixed costs any more -- you still stay in business because your upfront costs are now sunk. Getting into the business turned out to be a bad decision, but now that you're in it, leaving would be an even worse decision. In fact, companies do not exit until their operational value falls beneath their salvage value (the firm's next best value), and often that can become very low. In very stagnant industries with declining prices (old steel mills) it is common to see no new investment, money losing operations, but only gradually contraction. This is because a money losing plant, for a while, remains a better choice than closing the plant down.

Mom-and-Pop businesses are so tough because their salvage value is really low. If your family has run a florist shop for generations, what are you going to do if you close that shop down? All you know is how to run a florists. The next best option for people deeply tied to their family business is usually pretty grim, so they will hang in there long after strict business ventures have closed up shop. Venture funded companies are strict business ventures, so they compete with these guys very cautiously.

I don't know much about Dave, but I've been reading him less than I used to. I am have no idea if he feels jealous towards Blogger because Blogger is more famous than UserLand, although UserLand may have been there first. I don't care. And while I admire his enthusiasm about pushing technology forward (at least his own kind of technology), his seeming disdain for profit making would make me, as an investor, very leery about giving him money. Entrepreneurs walk a hard line -- they need to belive in their company to make it through the tough times, but they also need to remain clear eyed about the business and its performance. The point of business is to make money. Investors and employees can then take that money and spend it on all the indulgences and philanthropic causes they like.

Dave is free to think of his business as a Mom-and-Pop shop or whatever, but investors won't think of it as a business at all because it's not, it's a lifestyle choice. People are free to make whatever lifestyle choices they want, they just gotta pay for them. But I don't know if this matters at all to Dave since he has a gig at Harvard these days. Good for him.

Regarding Google's "evilness":

1) They've taken out a patent on search, and I ask "what took them so long". My memory may be playing tricks on me, but hasn't Google been the target of patent-infringement lawsuits by no-name search firms? The awful truth is that if you can patent algorithms, companies must patent algorithms, if only to protect themselves from extortion. This is all dead weight loss waste, and makes a travesty of the US Patent office.

2) Google is also beefing up their advertising. To be honest, the only online ads I've ever found useful are the sponsored links associated with Google searches.

3) I find claims that Google is a flagrant freeloader ripe coming from someone who uses Google without paying for it, and benefits from Google helping others find his site, also without compensating them for it. I'm sure these claims are sincere, but they also reveal something myopic about people's appraisal of their situations. Update Martin writes in, pointing out that his "freeloader" comment was a dramatic overstatement. So Martin does not think google freeloads, but points it out as a strain of thought.

4) I'm also unafraid of what Google plans to do in the future. Kottke's analysis of Google's real assets is very good, and I'm not sure if the paranoia I'm reading into it is real, or just projected since I found the article via a very pessimistic Dave Winer. Last I checked, it was easy to put up a website. If the quality of Google's search results starts to deteriorate, anyone can come in and offer better search. Oh, the joys of competition.

Monday, February 24, 2003

Deterrence vs Revenge

My old college roomie wonders whether DenBeste is serious about deterrence, or simply consumed with a lust for Vengence when he declares that "France and Germany must be crushed". Given my recent post on thinking about anger as a consumption good, it's something that's been on my mind lately.

In narrow game theory terms, there is no difference between a threat, a commitment, vengence, or deterrence. For contracts to be successful, the party that enforces it or collects the penalty for breaches must have some independent incentive to do so, otherwise it is in everyone's best interests to simply renegotiate terms. In other words, for a contract to be truly binding, both parties must also be interested in whatever the contract promises, otherwise the penalties are simply not credible. Suppose a struggling dieter convinces a restaurateur to penalize him $10,000 if he catches him eating eclairs in his restaurant. When the time comes and the dieter desperately wants eclairs, he can tell the restaurateur that the $10,000 fine is so large he will never violate the contract and so will never collect the fine, making the contract worthless. Instead, why don't they renegotiate? Perhaps the dieter would buy a round of drinks for the house if he could be released from his contractual obligations?

To make a threat, or commitment, credible, you need to convince the other party that you will refuse to renegotiate even if doing so is in your best interest -- that is, you will accept no money over some money, or some pain over no pain. After all, why would you need to commit to something if it made sense for you to do it all the time? You only need to commit if there's a risk you will back out. Deterrence, at its heart, is the art of crediblity.

There are lots of ways to try and make threats (or commitments) credible. Some people act crazy (like North Korea) and call the other's bluff. Others talk of higher principles they will stand by no matter what. Some folks just have a short fuse and consume a lot of anger, enabling them to follow through on threats. So when is retaliation driven by the cold calculation of deterrence, and when is it driven by the hot flash of anger? Unfortunately, one can only tell after the fact--if it's of net benefit to you it's deterrence, otherwise it's just anger. Stumblingtongue's image of the accountant with the very large axe is not so far off.

Sunday, February 23, 2003

Telcom fiasco

Michael Powell's move to roll back TA96 failed, when the FCC passed a convoluted set of measures that will do nothing to help the telco sector. The Telecommunications Act of 96 tried to create "service based competition" by forcing regional Bells to lease their lines to new entrants at low rates. This had the predictable effort of decreasing investment in physical plant because the Bells could not charge for it. The resulting fall in demand has eviscerated the telecom sector over the past few years.

Powell wanted to shift from service based competition to facilities based competition. So instead of Sprint, AT&T, and Verizon trying to get your local phone business, wireless ISPs, the cable company, the electric company, and the phone company would compete for your connectivity needs. Historically, the Faustian bargain between government and telco has been "you provide these services and we will protect your monopoly". Powell wanted the new bargain to be "enjoy your monopoly on the pipe to the home, but now everyone with a pipe to the home can offer whatever they want down it" believing that, in the packet switched world, any pipe can be a phone, TV, radio, etc.

TA96 was also very ambiguously worded, resulting in fierce litigation (which also kills investment dead). This new act is likely to be the same.

Public opinion on telco regulation is also muddled. This appropriately titled article in the same breath complains that 1) incumbents should not be allowed to charge prices for their services, 2) incumbents are not investing enough in new technology, 3) entrants, who don't invest in new technology either by simply resell existing services, are going to go out of business and 4) small businesses aren't protected.

IP and economics

Sorry for the delay in posts--I was away in warm but snowy Baltimore last week but am now back cold but snowless Chicago.

I have not read the paper by Boldrin and Levine that this article comments on, but it sounds like a mathematical treatment of how intellectual property intersects with production functions and various other standard ways economists think about firms and investment decisions. The standard economic position is that since ideas are non-rival (ie. they can be shared like smiles, not like spanners) without artificial scarcity by government fiat, there would be no incentive to create them in the first place. The paper builds models showing that things other than government fiat can create natural scarcity, allowing producers to charge for services and therefore have incentive to produce.

The brilliant Kevin Murphy of Chicago points out that Boldrin and Levine assume demand for IP is very elastic, and that record companies pricing CDs as high above marginal cost as they do argues against that. (Quick economic refresher: the monopoly price depends on how sensitive demand is to price. If demand goes way up when price falls a little, the monopoly price will be very close to the competitive [marginal] price. If demand stays the same no matter how much the price falls, the monopoly price will be much higher than the competitive price.) Robert Lucas (also from Chicago) argues that while government created scarcity might be valuable in some areas (drug patents) it seems not to be in others (music).

And I think that's exactly right. The Boldrin and Levine paper goes way to far asserting that IP protection is not necessary for innovation in all areas. The harms from IP protection are well understood by economists, but poorly by people in general (since they mostly have to do with dead weight loss from blocked transactions). The benefits of IP protection are poorly understood by everyone and philosophically driven by the lawyers representing the content industry, which is why we've seen copyrights extended retroactively, trademarks take on political dimensions in URLs, and patents covering software and business processes.

Since the Internet has lowered distribution and reproduction costs, bad IP laws are more costly now than they were in the past. 150 year copyright terms just didn't matter much before Napster. A new IP regime needs to understand the economics of production for different types of ideas and tailor the right laws for the right circumstances. Personally, I'd like to see shorter copyright, no patents on business processes or software, and longer patents for drugs.

Update: Arnold Kling comments on the same article.

Monday, February 17, 2003

Consumption taxes

This Slate article is a good example of how people don't like consumption taxes, or any taxes, because they are regressive. But there is more you want from a tax system than merely "progressivism" -- you also want it to be cheap. Whatever your position on how large government should be, it is clearly beneficial for taxation to be as efficient as possible (because this lets you tax more, or because this makes the tax burden lighter to bear, whichever you prefer).

Taxes become expensive when they 1) are applied to things that can avoid them and 2) become big (more precisely, when the marginal rate becomes large). Consumption taxes, whatever their regressive drawbacks, are more difficult to avoid than income taxes because there are more ways to hide income than consumption. They also have the benefit of broadening the tax base and letting marginal rates come down, and marginal rates, at ~45% are high. Given that the inefficiency (dead weight loss) caused by taxes increases to the square as the marginal rate increases linearly, reducing a high tax by even a few % can boost growth by a lot, benefiting everyone in the long run.

But in the long run, we are all dead. This post does not take a stand on what the right tax level is, but points out that its valuable to tax efficiently, and there are real tradeoffs between efficient taxation and progressive taxation.

Sunday, February 16, 2003

Modeling anger as a consumption good

One of the ways economics is weird is that it takes everyday words like "efficient" "rational" and "happy" and uses them to describe mathematical derivations. Of course, the results are muddled, and hilarity ensues. I also feel that this is why so many people feel such a revulsion to economics (when it's not boring them to tears) -- they have strong emotional ties to words that the cold mathematics of economics ignores.

My old school chum has not only a nifty new site design, but also a thoughtful post on how vindictiveness serves a useful role by acting as a deterrent. This is all very appropriate in light of the potential war in Iraq, but I'm not discussing that on this site, so let's model anger as a consumption good instead.

There was a sociological experiment where a bunch of strangers got to divide up money between them. If everyone played nice, they would all get the same amount of money. If someone played mean, he could get more money at the expense of the other players. In retaliation, other players could take money away from the cheating layer, but this cost them some money too. The way the experiment worked out in real life was that most of the time people played nice, occasionally someone played mean, and when they did the other players would get angry at him. There was usually one player who would get most angry, and after shouting, would start spending money to hurt the mean player, while the other players watched.

I think this was very revealing. Stopping cheaters makes everyone better off, but it also comes at some personal cost to yourself. An emotion like anger makes you not care about the harm you are doing to yourself as you teach that cheater a lesson. In that, through getting angry, you give up something to get something else, you can think of anger as a consumption good, like yogurt or Levi jeans, where you give up stuff (money) to get stuff (the pleasure of being angry). It is anger that makes any ex ante threat ("don't cheat or I will hurt you") plausible, because the cheater knows doing that will hurt you too. Please note: no one is being irrational, they're just consuming "anger" because, it makes them more "happy". (See what I told you about economics translated poorly).

Another key insight -- the other players who had been cheated got all the benefits of the cheater being punished without having to bear any of the costs. Yes, there is a freeriding problem with retribution, because we get to ride on the costly rage of others.

So, the two economic insights you get by modeling anger as a consumption good are 1) anger is expensive/keeping your temper makes you richer; and 2) there is not enough anger so more people cheat than would be optimal because not enough people tell them to stop.

Google buys Pyra

Oh oh -- it looks like content distribution (well, discovery anyway) has integrated with content production! Call the FCC!

Seriously though -- this may have ramifications for the powerlaw distribution in weblogs mentioned earlier. As I said in that post -- "[a]ny technology that strengthens non-weblog quality content location will weaken the power law distribution" and Google, last time I checked, was certainly a quality content location system that was not a weblog.

So, might some fancy Google collective algorithms weaken the powerlaw distribution? Maybe. Do we care? Actually, I don't think so. People who get hung-up on the inequality of blog readership aren't asking the right question -- how would the system respond to a shock? Suppose an A-list blogger goes bad, or stops blogging -- how long will they retain the #1 position? Or suppose a great new blogger starts writing -- how long will it take them to find an audience? If either of these take a long time then, and only then, is the powerlaw distribution inefficient.

Thursday, February 13, 2003

Varian on Taxes

Hal Varian is a good economist, and I recommend "Information Rules" which he co-wrote with Shapiro. My buddy SC points to this NY Times Op-Ed where Varian says:
A recent study by the economists Alan J. Auerbach, William G. Gale, Peter R. Orszag and Samara R. Potter, "Budget Blues: The Fiscal Outlook and Options for Reform," lays out the facts ([and concludes] that current patterns of spending and revenue are just not sustainable. Large future tax increases or drastic spending cuts are virtually inevitable.

...After 2012, the retirement fund surpluses shrink and eventually become deficits. According to the economists' projections, the spending on Social Security, Medicare and Medicaid will grow from 9 percent of G.D.P. in 2001 to 21 percent by 2075. "These three programs," the economists say, "would ultimately absorb a larger share of G.D.P. than does all of the federal government today."

...What will happen if nothing is done? If deficits continue to accumulate, the temptation to print money to pay our debts will become almost irresistible. Inflation is all too tempting as an "easy" way to avoid the political pain associated with tax increases or budget cuts.
Inflation, btw, is a tax on savings.

Mobile Phone + Camera Update

Bjoern Hansen has a followup to my recent post on SMS text entry vs photos. And yes, I too am lusting after a Bluetooth enabled phone I can synch with my Mac (right after I figure out how to get Now-Up-To-Date, iSynch 1.0, Address Book, and my Handspring talking to each other again).

Wacky Enron story

I find it difficult to parse this story about Enron. Apparently Enron engaged in a complicated tax regimes to avoid taxes on all those profits it was not making. I assume that, had it been honest about sucking wind, it would not have needed complex tax shelters.

Live music filetrading

Ages ago, I predicted that recordings from live concerts would be an important source of digital music online. I belive Daniel Berlinger had some issues with that, arguing that recording live events was costly too. It seems that for some kind of music, a good spot and DAT recorder gets high-enough quality sound and these recordings are being distributed online.

Ironically, the biggest impediment to this type of system gaining adoption is the RIAA's failure to get DRM technology "invasively and pervasively" deployed. One of the current proposals --a tax on ISP fees to make up for online file trading -- gives us the worst of all worlds because it 1) reinforces the RIAA monopoly (they now get to tax everyone with a broadband connection!) and 2) blocks alternative sources of music (because Britney Spears mp3s are still available for free).

Mobile Phone + Camera

I find SMS text-entry slow and painful, and belive it's popularity in Europe stems in large part from the high per-minute charges Europeans endure. But taking a picture with your phone and mailing it to someone is quick and easy, so I think it will prove quite popular. Here in the US, Sprint is offering an external camera attachment with some phones -- this will fail because 1) lugging a camera attachment around and then 2) plugging it into your phone before you take a snap shot is a Big Pain.

Do deficits restrict spending?

I took Brad DeLong to task for treating 50 year budget protections seriously (his reply here) and argued, as many do, that deficits are the only politically feasible way to cut spending, and that no, one does not need to be a baby-eating monster to think lower government spending is good.

Jane Galt has a super post on what the data says about expenditure vis a vis budget, and finds that it seems it have a small impact on interest rates. This makes sense because larger deficits require higher government borrowing and also reduce the likelihood of default, and while the US is no Argentina, you would expect this to feed (gently) into interest rates. Higher interest rates make borrowing more expensive, which increases the incentive to cut spending.

Wednesday, February 12, 2003


Clay Shirky wrote a really nice piece explaining why the blogging world has a few "stars" amidst a sea of low traffic sites. It turns out that when consecutive selections are influenced by previous selections, you end up with a "power law" distribution that has a few stars and a slew of, well, not stars.

Since this is an unequal distribution people complain because Unequal Is Bad, and then fixate on the equation that describes the distribution. It's worth noting that even in a zero-coercion environment (people picking which weblogs they want to read is about as non-coercive as you can get) the notion that a few writers win big while most people don't strikes folks as...sinister.

The power law exists in weblogs because you need weblogs to *find* good content--the best pieces end up on high-profile blogs because that's where the author wanted it to be promoted, and that's where the reader wanted to find it. Any technology that strengthens non-weblog quality content location will weaken the power law distribution.

Tuesday, February 11, 2003

Smart Mobs

When the government taxes one transaction to pay for another, it's called wealth "transfer". They do this both through explicit taxes, and through rules and regulations that act as implicit taxes. Currently in Chicago, for example, some legislators want to force real estate developers to set aside 30% of all new buildings to low-income housing. This requirement is identical to the government taxing new real estate development and using the proceeds to buy low income housing.

When there is money up for grabs, people will compete for it. This is called lobbying if done by corporate groups, or grass-roots activism if done by non-corporate groups (economists call this behavior "rent seeking"). If the market for favorable political transfers gets very efficient, then the money spent lobbying (or agitating) for these transfers will become equal to the value of the transfer itself. In this situation, there will be no net gain to those courting political favor, and the entire value of the transfer will become dead weight loss (to the detriment of the economy as a whole).

Some people think that the internet, and mobilephone+camera+blog will usher in a new era of grass-roots activism where People Power will rise again and become a potent political force. I don't know about that, but it will make the market for wealth transfer more efficient, resulting in smaller gains from lobbying. This will not result in less lobbying, as people will lobby until the gains for $1 of lobbying bring political benefits of exactly $1, it will just mean that more of the economy-neutral wealth transfer will be consumed by economy-destroying dead weight loss.

Friday, February 07, 2003

2050 Fiscall Projection

Prof. Brad DeLong comments on my recent post regarding the validity of 50 year projections:
50-year projections are hard, but not that hard. We have the demography fairly well understood. We expect further progress in medical science and further escalation of medical costs. The big uncertainties are (i) big wars, and (ii) the overall pace of productivity growth.

Those create huge variances around fifty-year projections. But that does not mean that such projections are not worth making. Indeed, the people I know who work on Social Security routinely call for projections out into the indefinite future.
He also had a thoughtful note about getting snarky in defense of/attacking certain political segments and the effect this has on one's reputation. That's good advice for everyone.

I want to mention cutting government spending. Taxes create dead weight loss by driving a price wedge between producers and consumers and thus preventing transactions that would have happened from happening. This slows down growth and results in less money for everyone. This effect also becomes geometrically larger as the tax rate grows, so a 5% tax does not create much loss, but a 40% tax creates lots (about 60x more than a 5% tax). This is a good reason to want lower taxes.

But how do you cut spending to afford those tax cuts? Public support for the Arts (consumed by the rich) is as vociferously defended as welfare (consumed by the poor) and every body feels that their pet program has to be kept no matter what. Politicians expand public services when the economy is doing well even though those services might be less in demand. In short, there is no political will to ever cut spending -- unless of course the electorate is knee deep in red-ink and don't want their taxes any higher. How do you create this red-ink? Deficit spending, baby! It's not pretty, it's not clever, and it's not cheap (you pay through higher interest rates), but the benefits may still outweigh the costs. I'd like to see better mechanisms to control and reduce spending.

Arnold Kling and Brad DeLong have more good posts on the subject.

Thursday, February 06, 2003

Et tu, Brad?

Apparently inspired by Krugman's heart breaking descent from thoughtful economist to raving partisan, the usually insightful Brad DeLong plays some of the same statistical tricks he usually calls other charlatans out on on his own readers, and perhaps, himself. The Terrible Thing the Evil Bush Administration Has Done Today? Made the country bankrupt by 2050.

2050. Twenty-Fifty. I will be almost 80! years! old! by 2050. Projecting any trend out 10 years is asking for trouble, but 50 years? This is so flagrantly dishonest that, in Brad's own words it really is beyond my comprehension why anyone would do this. Unless, of course, your thinking is impaired by personal bias.

Now, to be fair to Prof DeLong, it is human to have one's thinking impaired by personal bias, and Brad has made his leanings honest and public. The truth of the matter is that predictions are very difficult, especially of the future, and 50 year projections are worthless. It is true, however, that Republicans are using deficits to cut spending, but it is also true that many people are in favor of lower spending, there are good reasons for lower spending, and the US government does not have the equivalent of its monetary regulatory apparatus on its fiscal side, so it relies on deficits. Improving on that is, I think, one of the more interesting problems in macro right now, and I hope we'll have made good progress on it by 2050. (Yes, deficits can also be bad).

Tuesday, February 04, 2003

Pricing pressure on Microsoft

Microsoft has always been very conservative in its SEC filings, claiming that although things were good this quarter, it could all fall to pieces Any Day Now. It's latest pronouncement of impending doom (as it racks up ever higher profits) was the Open Source Software could force it to lower prices. In fact, they've already being doing this in certain government contracts where, I presume, lack of concern with productivity made the threat to move to OSS more credible.

Now I'm a big fan of OSS, but I also recognize that Linux will not replace Microsoft on the desktop anytime soon. What it will do is reduce Microsoft's ability to price because, even if it is not as good, it's good enough lower people's willingness to pay for Windows/Office.

Powell at the FCC

Here's a positive article from the Economist outlining what Michael Powell, now in his second year as chairman of the FCC, might have planned. They anticipate widespread deregulation in 1) media ownership, 2) last-mile telecom requirements, and 3) spectrum.

I personally am in favor of all three. While people are understandably suspicious of large media conglomerates, I think the economic argument for them censoring content is weak, although it's been helpfully pointed out to me that some sports games have been unaired because of a dispute over pricing. Also, the TELRIC pricing in the Telecom Act of 96 proved hostile to investment and impossible to administer, so I am comfortable giving back the ILECs last mile monopoly over copper, confident that cellular, wireless, and other new technologies will route around them. We'll see if this confidence is well founded. And finally, more open spectrum will be great -- I've been skeptical in the past, but the success of 802.11b shows what can be done if the unregulated spectrum is appropriate. And how can you not root for a guy who refers to his TiVo as "God's Machine"?

Sunday, February 02, 2003

No Blood for Oil

I've avoided discussing the potential war with Iraq, plenty of other places on the web for that, but I will comment on the "No Blood for Oil" rallying cry popular with the far-left. Their argument is that the US just wants to attack Iraq because then it will take over its oil fields and give itself oil for free. Whatever the merits of their claim, it doesn't stand up well to economic scrutiny.

Suppose you live with your elderly aunt, who does not charge you rent. Suppose then that your aunt passes away, leaving you the house in her will -- are you still living rent free? The intuitive answer "yes -- I was paying nothing before and I'm paying nothing now" also happens to be wrong because it ignores opportunity costs. When your aunt did not charge you rent she was, in fact, paying it herself because she decided to forgo the income she could have received by renting it to someone else. Forgone income is as much a cost as expenditure -- money you decline to take is as unavailable as money you exchange for something else. But when the house passes to you, and you decide to continue living in it, then you are paying the opportunity cost by not renting it to someone else.

In the abstract, people don't think about opportunity cost, but in real life they act as if they do. The person who inherited her aunt's house may decide to sell it and move somewhere smaller, and perhaps plan to spend the money she would have left over on a car. This decision is taking opportunity cost into account, although the term never conciously comes up.

If the US took over Iraq's oil fields and then gave itself the oil for free, it would bear the opportunity cost of selling the oil to the world markets. Given that oil is a commodity product, this price would be exactly what the US would have had to pay for it anyway, whether it was in Iraq or not. There might be other benefits (and costs) that come from controlling the oil, but "getting it for free" isn't one of them.