Broken Symmetry

Tim O'Reilly's Hymn for Open Government

A bit more poetic than usual, even for him:

There is a greater and greater gap between the systems used to manage complexity in the private sector and the same systems in the public sector. An average citizen carrying an iPhone has access to real-time "situational awareness" that outstrips military systems costing millions of dollars. From WalMart to Wall Street, from Google to the next generation of personal robotics, intelligent algorithms process massive amounts of data in real time, and respond automatically in close to real time. Meanwhile, government regulators too often rely on paper-based processes and reports that trail events by months, if not years.

...

The idea of building a platform, rather than building all of the end user applications yourself, turns out to be surprisingly relevant for government, allowing the private sector to build applications that the government might never have imagined, or been able to budget for if they did. As governments open their data and services to developers, we're seeing an explosion of innovation, and the development of new citizen services by the citizens themselves.

Unsustainable Growth, Startup Edition

From a post by Ron on things Paul Graham gets wrong:

It is true that VC returns have historically been bimodal, but this is not inherent in the nature of startups, it's because VC's engineer the environment to force the distribution to be bimodal. They do this because they are constrained by the structure of their funds to move money in and out on a particular schedule. They are usually not free to reinvest the proceeds from an early small win into a new startup that might produce another fast small win. Because of this, if a VC cashes out early, that capital can no longer work for them. So they use their influence to force companies to continue to grow past the point where they otherwise would have. This in turn forces the companies to take on additional risks that they otherwise might not have, which causes them in many instances to fail.

This argument is related to an earlier post on e the fundamental speed limit for organic growth. VCs are trying to force second-order phase transitions in markets. Sometimes that works, sometimes not. Organic growth is less risky, but also produces lower internal rates of return for its investors.

But I don't give VCs as much credit (or blame) as Ron for the outcomes. It's a transaction. Entrepreneurs have to take the money that VCs offer. Ultimately, there is too much money in VC relative to the number of scalable startup models that exist at any given moment in time because investors derive more satisfaction from seeing startups grow than they do from arbitraging currency or bond spreads (for example). We don't want to change that, that's a feature, not a bug. But it does suggest that taking other peoples' money to do VC investing is not a good idea for most managers.

These problems are working themselves out organically in the low-capital intensive consumer Internet markets. Paul Graham is partly responsible for that! Once the viability of new funding pipelines has been demonstrated, more capital-intensive industries will probably follow suit. But VCs in the more capital-intensive markets are already more patient, with lower expecations for exits, than the VCs in the lower-capital intensity industries.

Andy Grove on the Human Capital Needed for Scaling

This is easily the most insightful analysis of the most important issue the United States now faces that I have read this month:

Since the early days of Silicon Valley, the money invested in companies has increased dramatically, only to produce fewer jobs. Simply put, the U.S. has become wildly inefficient at creating American tech jobs. We may be less aware of this growing inefficiency, however, because our history of creating jobs over the past few decades has been spectacular -- masking our greater and greater spending to create each position.

My hope is that the loosening of the currency peg will help reverse some of these trends over the coming decades.

The Social Benefit of Banking

From the profile of Joe Evans in the WSJ:

"Real wealth accumulation in banking is the law of compound interest for long periods of time," he says, criticizing bank CEOs and directors who became obsessed with their stock price.

By extension, wealth accumulation that occurs faster than interest can compound on principal must be derived from pecuniary transfers -- or at least the transactions costs charged for those transfers.

One of the simplest ways to check whether growth of a process is sustainable is to look at whether a stock price is going up faster than the company's free cash-flow can compound interest, even at an exponential rate. If so, then the market is guessing. Maybe it's guessing right. But it's still just guessing.

An Obesity Epidemic in Venture Capital?

I'm glad Fred Wilson wrote this so that I don't have to:

In short, since I started investing in the web in '93/'94, I have invested in about 100 software-based web companies. And the success rate of fat companies versus lean companies is stark. I have never, not once, been successful with an investment in a company that raised a boatload of money before it found traction and product market fit with its primary product.

Fred's post is in response to a provocative piece by Ben Horowitz, The Case for the Fat Startup.

Fred does a better job than I could have taking this apart because of his superior experience. But should any entrepreneur be fooled about whether it's in their best interest to take more money from a VC? Of course folks like Ben want you to take more of their (limited partners') money.

I was going to pick apart Ben's post and show how his evidence and reasoning were incomplete. Most importantly, Ben talks about "lean" in a very loose sense that doesn't really get at what folks like Steve Blank mean by the concept. Running lean doesn't mean not spending money. Also, Ben's sample of N=1 doesn't mean anything. Anytime somebody uses a stock price chart as evidence for an argument, be skeptical.

But I hesitated because there are actually points Ben makes that I agree with. First, contrarian thinking is a good start for anybody making capital allocation decisions, including startup CEOs. So I agree with Ben's point that now might be a better time to take money than three years ago, for example. That doesn't mean it's a good idea in an absolute sense though.

I also agree with Ben's emphasis that it is the intellectual property rather than the business that most startups are building:

If you are a high-tech start-up, your value is in your intellectual property. Don’t stare at your spreadsheets so long that you get confused about that.

But a suppressed premise here is that the ultimate exit is acquisition by a larger company. If the ultimate goal were indefinite independence post-IPO, then developing IP without worrying about the IRR might still be the wrong strategy for some startups.

Fred's credibility with me again rises. It's nice to see a VC explaining to entrepreneurs how the game works from the inside. Aside from the guys at Venture Hacks, who are a kind of benchmark for this, I haven't seen that (at least not consistently) from any other VC bloggers.

UPDATE: Ben replies to Fred's comments here. Read Ben's reply for yourself. His reply did more to convince me that Fred has the right side of this one. Ben's reply basically reduces to pointing out that achieving product/market fit is difficult and uncertain (okay, but how does having more VC money early make it less difficult and uncertain?), and that his N=1 sample is still representative of the population. The latter point is ridiculous. Even if he hadn't raised a bunch of money already (making the earlier stage VCs allies in raising more money), Marc had brand name at that point from Netscape that almost nobody has. It didn't matter that Marc was at Ning. Ben's failure to acknowledge this is disappointing.

Okay, Sornette. You have our attention. Now show us the Money!

The Physics arXiv blog today recaps the prediction of Didier Sornette and crew that the Shanghai Composite was a bubble that would pop between 17 July and 27 July. It didn't. It popped on 4 August instead:

Thanks to the Physics arXiv blog for putting together the graphic. I was bemused when I first saw Sornette's paper. See earlier post on the prediction here. The Shanghai Composite indeed looked a little overheated. But nobody can see into the future, and (more importantly) Sornette gave us almost no information about how and why he was predicted a crash within this window. Nevertheless I was impressed with his courage to make a prediction that is so easily falsifiable. Of course, this high-risk, high-reward strategy is also a favorite of cranks.

That being said, the general idea that super-exponential growth is unsustainable is one that I have explored myself on this very blog. See excerpt from this earlier post on e: The Fundamental Speed Limit to Startup Growth.:

the very fastest that you can grow anything is by continuously compounding -- by feeding whatever new revenue you generate back into generating more revenue instantly. If you plot revenue as a function of time, and look at the base of the logarithm, the largest base results when you instaneously begin compounding every new bit of revenue as soon as it comes in. That base is e. There is no way to grow revenue faster than e.

Although these statements are unequivocal, I will be the first to admit that it's very difficult to verify any of them empirically by looking at actual firm growth rates -- much less the Shanghai Composite.

So here we are, with a falsified prediction and no information about how it was produced -- only tantalizing hints. Okay, Sornette. For better or worse, you have our attention. Let's see how you came up with your prediction so that we can understand whether this was simply a lucky guess.

Economics that can be Falsified

From K. Bastiaensen, P. Cauwels, D. Sornette, R. Woodard, and W.-X. Zhou.

Amid the current financial crisis, there has been one equity index beating all others: the Shanghai Composite. Our analysis of this main Chinese equity index shows clear signatures of a bubble build up and we go on to predict its most likely crash date: July 17-27, 2009 (20%/80% quantile confidence interval).

Available here. Found via Physics arXiv blog. Can a reader tell me who authors and edits this blog?

Didier Sornette's work on the endogenous and exogenous trends in YouTube video watching was covered in this earlier post.

The argument here rests on the observation that faster-than-exponential growth is not physically sustainable. A similar point was made in an earlier post on this blog regarding venture capital financing and startup growth: The Fundamental Speed Limit on the Growth of a Startup.

UPDATE (August 2, 2009): Make that, economics that has been falsified. The Shanghai Composite blew through the intervaal identified showing few signs of slowing. Conway and Kochen trump Sornette et al.

Highlights from the 2009 IP Business Congress

Monday and Tuesday of this week I attended the second annual IP Business Congress, which this year was venued at the Four Seasons Hotel in Chicago. There were almost 400 attendees at the event, which was a good turnout given the economic climate. Attendees represented stakeholders in nearly every part of the IP Market, both primary and secondary. As expected, the panels organized by IAM magazine were excellent and informative, especially the question and answer sessions afterward. This is an event that I expect to put on my calendar annually.

Highlights for me as an attendee to this event included (1) the first panel on Monday on the state of IP rights in China, India, Japan, Europe, and the United States; (2) the panel on patent pools on Monday afternoon; (3) the panel on open innovation on Monday afternoon; and (4) the Gala dinner on Monday night. Tuesday's panels were also excellent, but of less interest to me because of my more direct involvement with many of the issues addressed on Tuesday.

State of IP Rights Internationally

Harvey (expert on China) kicked things off by noting how IP rights had been a strong influence on the development of rule of law more generally in China: IP was the first private property right to be recognized by the Chinese government after the cultural revolution. Harvey also explained that in number of lawsuits, there is more IP litigation in China than in any other nation. The picture he painted was of a thriving IP system in China that was operating decoupled from the rest of the IP world. The coupling of China's IP system into the rest of the world's will produce winners and losers of epic proportion, and now is the time for us to think strategically about how to prepare our clients and customers for this inevitable merge of the Chinese system into the rest of the world.

McGinley (EPO Officer) followed Harvey with sobering statistics on the patent application bubble that is growing inside the US, EPO, and Japanese patent offices: 2 million applications althogether are now pending inside these offices. Given how the rate of examination has not kept up with the rate of growth in filing, McGinley suggested that the system was beginning to resemble a pyramid scheme! This was a brave position for a government official to take, but everybody in the system should appreciate what McGinley is trying to do. Where might we have been had more financial regulatory authorities here in the U.S. pointed out the pyramid-scheme-like growth of the subprime mortgage market? McGinley was short on answers as to how to solve the problem, but he did suggest that Chinese officials were probably at the top in terms of training practices and the reduction of systematic errors in examination. This seemed unsurprising to this listener given the on-the-ground experience of many Chinese with more complex manufacturing workflows!

The presenter from Japan had little of interest to say, and presented some confusing interpretations of statistics on patenting in Japan. For example, he pointed out that the largest number of patents issued in Japan were to computer companies rather than pharma companies. Pharma companies need only one well-drafted and prosecuted patent to protect their products. This is not very surprising, and is true everywhere in the world.

Basheer from India gave in many ways the most interesting presentation. His was mostly a diplomatic presentation of facts about the IP market in India, where the political agenda is driven by the success of generic drug manufacturers. As a few BS readers may know, Basheer is also one of the authors of the consistently excellent (but very detailed!) Spicy IP blog. From listening to Basheer, one could piece together a picture of India in a critical state, in which more or less IP protection could result from a sharp push in either direction. Too difficult to call for the time being. On my view, it seems that Indian biotech and pharma researchers have a regulatory arbitrage over their U.S. counterparts, who have many years of redtape to cut through before carrying out even relatively safe clinical trials! Note that beating heart transplants are now being carried out in Indian hospitals. There are many high-risk, high-reward medical treatments now avaiable in India that may never be available in the United States. Another trend to watch carefully.

The first question/comment for this panel was also notable. It came from Sherry Knowles of GSK, who noted that Chinese generic manufacturers were getting drugs to market before US manufacturers by simply scraping clinical data off the clinicaltrials.gov website on which U.S. manufacturers, such as GSK, have to post their data by law! There's one for the USTR to call to Chinese government attention.

Patent Pooling

The Monday afternoon panel on patent pools included Sisvel founder Dini, Hinman from Verizon, and Michel from the FTC. Dini gave a great pitch for hiring independent administrators (such as Sisvel) for allocating profits among patent pool participants, and for getting such admins involved at as early a stage as possible. Michel gave a presentation of the recent history of FTC regulation of patent pools, including a summary of the federal court holdings relating to Rambus's participation in JEDEC. Hinman talked about the perils of patent pool participation, and noted that even opt-out could be anti-competitive. Referring to one patent pool in particular, he pointed out that all but two industry participants had entered. I saw Michel scribbling notes!

The upshot is that if your company is involved with patent pools, you need to hire experienced outside counsel with knowledge at the interface of IP and Antitrust law to help ensure your business strategy reduces the risk of antitrust liability. Your author knows and works with a few such experts, and is happy to make an intro. Note that most corporate lawyers or IP lawyers will not be aware of the perils of competition law because the Bush and Clinton administrations were relatively lax on this front. All indications are that the Obama administration will be more active in enforcing competition law.

Open Innovation

By now the term "open innovation" has become relatively well-known within the business community. Some IP owners may believe that open innovation represents a form of socialism -- an alternative to a system of property rights. Attendance to this panel would have been the antidote to such misunderstandings of this theory. Panelists Weedman from P&G, Simonton from Exelixis, and Bergelt from the Open Innovation Network (and formerly of Paradox) talked about the dynamic growth that the feedback loops established by open innovation networks facilitate for the companies that take open innovation seriously.

Because the theory of open innovation is a paradigm shift from traditional neoclassical economic theories, only a handful of businesses (including P&G and Exelixis) have been fortunate enough to find leaders willing to foster the organizational commitments necessary to make open innovation work. This business model is the future of every large public corporation in the world. Watching these businesses continue to outgrow their competitors, others will eventually follow.

Gala Dinner

One of the best parts of the event was meeting the attendees. Over drinks before the Gala dinner, I had the chance to catch up with Niels Reimers, who later that evening was inducted into the IP Hall of Fame for his path-breaking work in technology transfer. In this rare moment of opportunity, I had the chance to ask Niels about how he worked at Stanford. By Niels's description, he was practicing the same technique followed by the best founders and managers in Silicon Valley both then and now -- "management by walking around." Dave Packard and Bill Hewlett, Bill Perry at ESL, and Nils at Stanford were all doing management by walking around. Niels's personal accounts of his work connecting Stanford music department (!) professor John Chowning with Hamamatsu corporation engineers were a highlight among highlights for this event.

Later, when Niels cited the Teece Model of innovation in his acceptance speech, I nearly fell out of my chair. The same model was cited by FTC Commissioner Tom Rosch in his speech about new theories in competition law on June 1, 2009. Note that Teece was also cited in the explanation of systems theory provided in your author's letter comment to the FTC on the Evolving IP Marketplace.

Final Thoughts

Another highlight made possible by my attendance to the event, but not directly related thereto, was an invitation to a breakfast with Congressman Bill Foster on Tuesday morning. I am already a fan. Foster is an entrepreneur and physicist who decided (after 22 years of building superconducting magnets to detect subatomic particles) to roll up his sleeves and try to help fix government. I'm pretty sure there's nobody else in Congress who would know what the title of this blog refers to. The discussion of the patent system and innovation at this breakfast was the best I've heard ever anywhere, but it also revealed to me the underlying problem facing patent reform: nobody has a complete view of the whole system. Almost everybody (at least) focuses on the broken incentives and procedures facing them in their subsystem, without stepping back to see how all of the subsystems coalesce into an competitive and innovative ecosystem. Nonetheless, the kinds of questions asked and answers given at this breakfast suggest that already the level of discourse has progressed far beyond the pat answers that were served up last year when patent reform was before Congress. There is more widespread awareness at least of how the public has been manipulated to some extent by the more technologically sophisticated and media-connected subset of stakeholders in the system. Perhaps we will have sensible patent reform after all. Here's to that hope.

e: The Fundamental Speed Limit on the Organic Growth of a Startup

Fred Wilson had a couple of good posts on the Venture Capital Math Problem last week. I made a few comments over there, but I wanted to offer a few more extended thoughts on this problem.

Although I haven't found anything on this in the literature I've read on finance, it is well known in other fields, including physics and biology, that there is a fundamental speed limit to the rate of growth of any quantity -- namely, e.

There are many places where one can read about e on the Internet, so I'm going to confine my remarks to the question of how this fundamental speed limit is relevant to venture capitalists and entrepreneurs. For one good introduction to e in finance, see here. But for one tantalizing hint at how interesting things get when one starts looking carefully at e, note that because most quantities we are interested in measuring vary in time (i.e., they have a relative-phase that is a coefficient to the complex quantity i), e together with pi and i provide a fundamental connection between algebra and calculus, or (equivalently) between geometry and dynamics.

Entrepreneurs and the venture capitalists who finance them want their businesses to grow as fast as possible. At the nitty gritty level, this means building a product or service and creating a market for that product or service as fast as possible. In practice, this process of development, marketing, and sales, occurs in cycles. If things work right, with each iteration of the cycle, the product or service is adjusted to make a better fit to the existing or forecasted market, the market then expands either by word of mouth or by direct advertising, and the process repeats. Taking also into consideration the fact that as the market expands beyond a certain size it will begin eating into the money that used to be spent on other similar (but not identical) products or services, and you've got what Joseph Schumpeter called waves of "creative destruction."

Imagine you had the best product idea and the best team in the world. What is the fastest that you could possibly grow your startup? Would pumping more money into the startup help? Every venture capitalist would like to know the answers to these questions.

Some already know that the answer is that the very fastest that you can grow anything is by continuously compounding -- by feeding whatever new revenue you generate back into generating more revenue instantly. If you plot revenue as a function of time, and look at the base of the logarithm, the largest base results when you instaneously begin compounding every new bit of revenue as soon as it comes in. That base is e. There is no way to grow revenue faster than e.

In fact, you won't come close to compounding continuously most of the time. That's because there is always going to be some dissipative force in your organization that drains revenue off to processes other than the ones that are iterating and expanding your market share.

As a final thought, notice that if we assume a growth rate equal to f(x) = x^k (a power law), then log f(e) = k and log f(1) = 0. In other words, for our growth rate to scale up to new market sizes perfectly (that's what taking the logarithm requires) we have to be continuously compounding, which I just said we can't do. No startup's rate of growth is going to scale perfectly. The best one can hope for is to be closer to e than to 1.

UPDATE: It's worth pointing out that under the right conditions, change can happen faster than exponentially. Kurzweil has his double-exponential theory. Power-law fanatics have theirs. I can see how this might work by analogy to phase transitions in condensed matter physics. The organic growth described above would apply to a continuous transition. In some cases, however, there is a certain amount of energy that disappears from the system as local symmetries are broken -- like the latent heat of fusion as water goes from liquid to solid. In certain cases, one might be able to supercharge growth by adding a lot of resources into building out a basic structure that can scale with logarithmic increases in demand. But that would seem to be a risky proposition compared with organic growth in many scenarios.

Feedback Loops in Urban Planning

From Seed Magazine, a conversation between Steve Strogatz and Carlo Ratti:

CR: I like your example of the power grid. But isn't part of the problem that we're treating new distributed systems with old, centralized controllers? What if the intelligence were spread into the system? Then the system would regulate itself in a different way.

SS: Right. That's actually the solution that nature tries to use. So, the thing is if you're going to create these big, complex networks, you better be prepared to control them in a decentralized way also. The two should go hand-in-hand. In the case of the financial system, we don't have anything like that really. We use the language that comes from the power-grid scenario. We talk about pulling the circuit breaker when the market has a problem. It's a very crude kind of response.

[snip]

SS: At least in chaos and traditional nonlinear dynamics we had pictures. We could understand what was happening by visualizing these pictures in two or three, and sometimes four or five, dimensions. But now that we're dealing with network systems with millions of nodes, we can't use geometry either. We need something else. We have graph theory, that's something. We have simulation but the simulations are often as hard to understand as the reality. So I think we have a real psychological question here as to whether we will ever be able to develop what's needed to understand this modern world. And it may be, speaking of evolution, that we're not the ones who will understand it. The artificially intelligent devices that we create will understand it and they will report back to us. I think that's probably the most likely scenario. If we survive this networked world it will be because entities much more intelligent than us will figure out the laws. Who knows? The history of science has always kept moving forward, but, on the other hand, that doesn't mean it always will.

See earlier posts on Strogatz and Sync here and here.

Maybe I'm being too optimistic here, but I think geometry hasn't exhausted it's usefulness to understanding nonlinear phenomena. In particular, I think complex analytic geometry is still in its infancy as a field.

Underappreciated Cause of the Crisis

The same rapid technological advances that are driving the marginal cost of information dissemination and trade execution to zero may also have implications for the incentives faced by information producers. As trading costs fall, the marginal investor may be becoming less experienced, less sophisticated, and less able to derive fundamental security values from raw information. At the same time, the characteristics and business models of publicly traded companies are changing as the economy itself changes. Newly listed firms are far less likely to generate sufficient cash flows in the near term to internally finance the costs of expansion and compensation. A high current stock price becomes a vital cash substitute for these firms. These trends combine to create strong incentives for firms to distort the information they produce to the investor community. We show that, indeed, there is evidence of deterioration in the quality of information that firms supply to investors. This deterioration may slow down the improvements in security markets that technological progress brings about. Moreover, private mechanisms such as litigation and information packaging by financial intermediaries are unlikely to solve the problem of information quality. Investor education and the regulation of information disclosure in particular may therefore play an important role in the development of securities markets in the foreseeable future.

From Technology, Information Production, and Market Efficiency by Shleifer et al.

Here are few interesting charts from the paper:

The units for the y-axis here are percent of average firm assets.  Discretionary accruals are basically adjustments to cash flow permitted under GAAP.

Is there any correlation between how many discretionary accruals are used and future returns?  Yes.

Accounting rules need to be made simpler.  Corporations must be required to report more information, not less.  Intermediaries who process financial statements and other accounting information produced by publicly traded corporations need to be independent of the investment banks or other insitutions that finance publicly traded corporations.

Would you still read Consumer Reports if it were paid for and published by GE, GM, and Dell? 

Are Fannie Mae and Freddie Mac evidence that Creative Capitalism cannot Work?

Not quite.

Over at the Creative Capitalism blog, Lawrence Summers points to the failure of privatization of government home mortgage lending as evidence of the limits to Creative Capitalism.

But the problem with these institutions, as Summers seems at least implicitly aware, was not with the existence of public/private partnerships, but with the bundling of incentives effectively implemented by these partnerships.  Specifically, with the government put on F&F mortgages, the banks buying them could afford to ignore low probability, large-scale catastrophes.  The ability to ignore this contingency, ironically, is part of what led to its more rapid development.  Markets are cyclical.  A clump of clothes in your washing machine can throw the whole machine out of whack.  How much more so a clump of poorly priced assets in a market?

The way forward, as I blogged a few months ago, is to focus on how assets are bundled.  Within any market, the smallest unit of demand must both receive benefits and pay costs of social consequences to a transaction.  In the mortgage market, this means that home buyers should also have been paying for insurance premiums for the risk of price declines due to offsite factors.

If we give up on creative capitalism now, we'll only be back where we started with marxism vs. laissez-faire.  We shouldn't give up just because it's hard.

This is not healthy for lawyers or clients

What you're looking at is the bimodal distribution of starting lawyer salaries.  If you integrated up the area under the right lobe, you'd figure out pretty quickly how many lawyers are working in "big law," and if you compared that number to the number graduating from top-U.S. News-ranked law schools, you'd get a pretty good idea of where in the U.S. News rankings your chances of making the big money fall off dramatically.

This distribution can be interpreted in a number of ways.  But perhaps the simplest explanation is that the demand for legal services from large corporations is concentrated on the supply for a handful of top law firms.  Why is this?  Because it's very difficult to measure the quality of legal services.

How could you change this?  For starters, you could change the rules of professional ethics that prevent lawyers from partnering with clients to provide services.  If lawyers and clients could share equity, their interests would be more aligned, and billable hours and pedigree would be less necessary as proxies for quality.

Via Greg Mankiw

Creative Capitalism = Sustainable Capitalism

I've been enjoying the commentary posted at the new Creative Capitalism blog.  I love it when people at the very top of their game get together and talk in open-ended terms about how things could be improved.  Being able to listen in -- and even participate through comments -- is a totally earthshaking technological change.

I'm on Peter Drucker's side of the discussion about creative capitalism, and whether it has any instrumental value to investors and managers of large institutions.  Here's what Mr. Drucker wrote in Management: Tasks, Responsibilities, Practices:

In modern society there is no other leadership group but managers. If the managers of our major institutions, and especially of business, do not take responsibility for the common good, no one else can or will.

Drucker knew that the managers of private institutions would have to take more responsibility for the common good because he understood the limits of centrally planned institutional design.  Decentralized control through sponatenous ordering is the only sustainable form of institution over long time periods and large geographical areas.  The administrative agencies of the Executive branch of federal government in the United States are an example of how central planning cannot keep up with markets (just look at how well the Federal Reserve and SEC have kept up with the markets).

But how can we improve on the model of management and investing followed in the United States?  We have to go back to the fundamentals, and rearticulate why it is that we're measuring and reporting the things we do.  In particular, I believe we need to make some important adjustments to how we report things on financial statements.  If we were all measuring growth in the same way, then cooperation among the various stakeholders in corporations would be much easier.  And competition among various corporations would be more efficient.

So here's my suggestion:

For every balance sheet account, let's include a measure of the average debit flow and the average credit flow in/out of the various accounts on a daily basis over the time period of the financial statement.

Thoughtful managers and investors may recognize how this additional measure of liquidity would help us to distinguish between a healthy corporation that is on track to grow sustainably during the next period, and a sick corporation that is being unwound slowly to generate as much cash as possible before dissolving.

Consider who within our economy is now best trained at measuring and reporting the average frequency of different economic events.  With that answer in mind, is it any wonder that some of the best investors in the world generate new liquidity through insurance float?

Here are a few links to other posts I've made about how we could do accounting and management differently:

• Inventors are customers
• Markets and firms maximize profit by synchronizing the flow of supply to demand.
  • The neuroscientific mechanism for synchronized flow.
  • Relationship to Coase, Alchian & Demsetz, and Williamson.
• Could Bear Sterns have avoided disaster were debit and credit flows reported to its management?
• Cost accounting and lean accounting are both needed as views of firm financial health.
• Why Wall Street ignores balance sheets
• Response to Summers on Regulatory System Design

How Bear Sterns is like Easter Island

Yesterday I got my hands on a copy of the latest book by Jared Diamond (author of Guns, Germs, and Steel).  His newer book, Collapse, is an anthropological study of how various societies around the world have failed through lack of coordination.  I couldn't help blogging the thought that some of the disasters he describes in this book could have been avoided had more information been available about the relative frequency of supply and demand.

In the book, Diamond describes how the resident of Easter Island disappeared after failing to coordinate consumption of the trees growing on the island.  They used them all up before new ones had time to grow.  The society collapsed afterward.

Imagine that everybody on Easter island could easily access the following information:

• the number of trees
• how many trees were cut down per day (or month, or year)
• how many new trees grow per day (or month, or year)

Even the worst mathematician could have recognized at a glance that the trees were going to be used up faster than they were being grown.

But are we really any smarter or better coordinated than the residents of Easter Island?  I say no.  For what's the difference between keeping track of the amount of subprime mortgage assets you have on your books without keeping track of influx (i.e., number of new subprime mortgage assets added per unit time) and outflux (i.e., the number of subprime mortgage assets sold per unit time).  In this sense, managers at investment banks like Bear Sterns didn't do much better than the residents of Easter Island.  Bear Sterns did do better at least in the sense that there are a few people left at Bear Sterns.  But that's cold comfort.

The moral of the story is that bad things happen when you don't keep the rhythms of supply and demand locked in phase and frequency.  Requiring companies to report turnover rates for each balance sheet account would be a good start in that direction.

Summers on "Regulatory System Design"

The Financial Times today prints this column by Larry Summers.  It's the most insightful thing I've read thus far on how the public regulation of the financial system might be improved.  In short, it's because Summers, unlike so many other commentators, has in mind a dynamic mental model for markets.

He gives six goals for public regulatory authorities.  I explain how each may arise from an insight provided by dynamic mental models.

1. Centralize decisionmaking authority over discrete markets or firms within one agency.

Over time, firms will shift from one way of framing deals to another in order to avail themselves of whatever set of regulatory rules is most favorable at a given moment.  So long as the substantive regulatory goals are shared among the agencies, these shifts merely raise transactions costs, redistributing wealth from the parties and counterparties to transactions to the croupiers with the connections and information to stay one-step ahead of the pack in predicting regulatory change.

2. Self-regulation is deregulation and supervise balance sheet data.

The first part isn't a dynamic insight.  It's the fundamental hypothesis of rationality.  The second part is, because the balance sheet encodes some information about the periodicity of assets, i.e., how often they turn over.  But making that information more explicit in financial statements would be an improvement.

3. Regulation must be premised on inability to predict the future -- i.e., on making the economy shock resistant, not on avoiding shocks.

Time figures in explicitly here.  Control theory is all about building systems to be dynamically stable without human control.  It can't be done perfectly.  But it's better than watching the system and trying to jump in when something goes wrong.

4. Regulation should be focussed on building a healthy system, not on punishing particular individuals.

This is related to the last point because the size of the population to which regulations are applied and the time horizon over which regulations are applied have similar consequences for the dynamic picture of firms and markets.

5. Regulation should acknowledge and address the possibility of multiple firms engaging in the same activities in a massively parallel way.

The frequency of supply cycles grow nonlinearly when multiple firms or people engage in the same supply activity.  This can amplify supply too much for demand to keep up.

6. Regulations should avoid letting individual failures create systemic risk.

This one is a bit inscrutable, but his gloss on it suggests that he's reaching for a principal of least authority -- i.e., that regulations should be devised to minimize the amount of affirmative intervention required by regulators.  Electrical engineers and physicists developed control theory to help them do just that.  No human being could fly the stealth fighter.  But with analog and digital circuits adjusting outputs to inputs almost instaneously, the stealth fighter can fly in a dynamically stable equilibrium.

As my old physics professor used to say, "YOU are not in equilibrium.  And that's a good thing."

Applying Control Theory to the Subprime Mortgage Supply Cycle

In an earlier post, I explained how transactions costs, liquidity, money supply, and scarcity can be modeled as the resistance, inductance, capacitance, and voltage of an RLC circuit.  When Schumpeterian creative destruction causes a demand voltage to spike, the coupled supply circuit will damp the spike at the "resonance frequency."  Oscillation will still be observed if the demand spike is large enough, but the resonance or bubble effect will decay to zero over longer time periods unless an active element is included in the supply circuit.

To show how the resonant frequency for this RLC filter can be calculated, let's apply some actual numbers to the supply cycle for subprime mortgages.

Because of MBS, CDOs, and CDO^2s, the liquidity in the subprime market was actually much larger than ever before.  How many orders of magnitude is hard to say.  I'll estimate that the rebundling and sale of mortgages increased the scale of supply by a few orders of magnitude, and also increased the number of days / sale as the bundling, rebundling, etc. takes longer to accomplish than a simple mortgage.  If it takes 100 days to do the bundling, but by doing so you get up to $10,000,000 scale, the measure of liquidity becomes:

L = (100 day / sale) / ($10,000,000 sale) = 1 x 10^(-5) days / $

Money supply (capacitance) should be measured in units of days / $ -- i.e., the number of days that each $ loaned from an external supply spends in the market before being paid back.  I'm going to estimate that about $1,000,000,000 of the entire money supply (on average) was locked up in mortgages, MBSs, CDOs, or CDO^2s.  And banks were probably holding onto these funds for about 100 days before going back to the Fed.  Of course the government kept messing with money supply by increasing and decreasing the cost of lending, but I'm ignoring that.  I'll repeat, however, that if the money supply obeyed something like Taylor's rules, we'd have a more determinate system.

C = ($1,000,000,000 * 100 days) = 1x10^11 $-days

Solving the differential equations for the frequency of the resonant peak for this cycle, we find that:

Resonant frequency = 1 / square root (L * C) = 1 / sqr ( 1x10^-5 days / $ * 1x10^11 $-days ) = 0.001 cycles / day.

In other words, according to these estimates of liquidity and money supply, we can predict a peak in the subprime mortgage market to occur about once every three years.  And that's roughly how long it seems to have taken for us to reach a peak once bundling of mortgages started.  In general, when you hit a damped harmonic oscillator (like an RLC circuit) with a signal spike, it will ring at a particular resonant frequency.  The resonant frequency is like the "sweet spot" on a baseball bat, whereby  an incoming spike at that exact frequency will be damped to zero at exactly the length of one period of oscillation.  If the demand spike has a longer or shorter frequency, then the damping will either occur in less than the period of the resonant frequency (overdamping) or the spike will continue to oscillate but at lower amplitude (underdamping).

When the liquidity and money supply effectively underdamp the incoming demand spike, the RLC supply circuit will continue to oscillate at the resonant frequency, but at successively lower amplitudes.

The Most Important Role of Public and Private Regulators?

Libertarians struggle with the question of how and when any government or private regulation is proper.  In the wake of the subprime mortgage market collapse and in view of the existing problems with the patent system in the United States, I have come to a few more general conclusions.

The most useful role for regulators is to ensure that the consumers whose transactions constitute the smallest units of demand within a market are required to internalize at least part of both the positive and the negative externalities (i.e., social costs and benefits not otherwise priced into the bilateral transaction) of subsequent comparable transactions.  This will require at least the cooperation of the individual or firm that is counterparty to the smallest unit of demand.

In the housing market, a bubble developed because at each stage in the cycle from homebuyer to bank, to larger institutional investors, and back to buyer, no accounting was made for how an increase in home prices was going to increase the price and decrease the volume of transactions in other goods.  In other words, the opportunity costs of increasing prices in homes was not sufficiently internalized by homebuyers in the subprime mortgage market.  This problem could have been avoided by requiring homebuyers to buy insurance covering the risk of declines in home price due to off-site factors.  (Thanks, Lee Fennell.)

Taxes are probably the most impactful way that government regulators have ensured that the negative externalities of an individual engaging in particular activities are internalized by the same individuals in some measure.  (Thanks, Pigou.)  In the sense I'm thinking about them, taxes are like an insurance premium we pay to the government to ensure that somebody is worrying about (and will bail us out from) major catastrophes that are too large in scale or too long in time horizon for the market to accurately price into transactions with consumers.  Please note that this does not mean that I am in favor of raising taxes!  A corollary is that the role of government may be rationally more limited in places and times when private institutions are able to more accurately price low-probability and long-time horizon (i.e., major) catastrophes.  Requiring insurance and reinsurance against catastrophes is probably healthier for everyone than building labyrinthine bureaucracies (like FEMA) that aren't going to do much work on preventing disasters, and won't be much help when they inevitably occur.

Here's one piece of evidence to support this theory.  Warren Buffett and Charlie Munger have almost made a religion out of including negative feedback in every financial statement, probably even in every transaction.  The result?  Longer time horizons and more accurate pricing (through negative feedback) have permitted them to run what is perhaps the most efficient conglomerate in the history of humanity.

How is it organized? I don’t think in history of world has anything Berkshire’s size organized in so decentralized a fashion. Net amount of bureaucracy is tiny, costs are low, autonomy in subsidiaries is vast, no common culture shuffling people around. How far can this go? This system has gone farther than any other system. Low cost, not a lot of envy effects – where everyone compares everything. People in subsidiaries have a feeling – whereby there is less fealty to headquarters. If you want an imperial headquarters which exacts a big overhead charge on the provinces – they will resent it. Net number of intra-subsidiary transfers is tiny. It has worked well. It can go a lot farther. No one else has been here before.

Berkshire-Hathaway is one of the first private regulators to make negative feedback a priority, and look at what they have accomplished.  What would the world look like if we redesigned legal and social norms to encourage negative feedback (i.e., accountability).  And Berkshire-Hathaway did it without any government telling them that they had to.  They were seed crystals in this sense.

Update: The characteristic slope of the Taylor Rule for monetary policy (which requires that the magnitude of changes in target rates always exceed the magnitude of changes measured in inflation) could be viewed as an embodiment of the negative feedback principle.  A control theorist might say that the Taylor Rule leads to more stable economic cycles by requiring a decrease in money supply in response to increases in inflation.  Assuming the measurements were accurate, in the limit in which the lag in time between the decrease in money supply and the increase in inflation approches zero (i.e., assuming adjustments could be made instantaneously) the money supply rate and the rate of inflation should be equal.

Thus, the best known approach to matching the money supply rate to the real rate of inflation is not unlike the approach followed by Buffett and Munger in their attempts to match the price of Berkshire-Hathaway stock to their estimates of its intrinsic value: negative feedback works.

Stable Market Design with Control Theory

Earlier this week, I had a vision of how analog circuit design theory could have provided some useful insights into how to avoid bubble markets.  I haven't found too much on this from googling, although this paper looks pretty close from the abstract (I don't have a subscription so can't verify whether they're actually thinking the same way).

The field that physicists and applied mathematicians call Control and Dynamical Systems has basically developed to aid engineers in building systems that use feedback to stabilize the state of any system that goes through cycles.  There are lots of things that machines (like the stealth fighter) do that humans would not be able to do because of the magic of feedback-stabilized oscillation.

The implications that this has for business cycles in public and private markets is so obvious that I'm quite certain that somebody already knows how to do this.   Alas, they're probably making boat loads of money on it as we speak.  Another problem worth solving is how to give incentives to such people to share their insights through something other than bidding or asking price.  (Actually, granting patents on financial engineering innovations isn't a bad way to do this.  But a two-decade term would be overkill in most cases.)

In the interest of aiding translation between physicists and economists (and hopefully help avoid yet another major bubble in our financial markets), I'm going to identify the simplifying assumptions that I think are most useful in modeling markets with control theory, and then offer a few extremely crude observations about the potential benefits of applying this theory.  I'll use electrical engineering terminology to show the relationship between the variables.

* Price can be modeled as a two-dimensional current signal in time and demand P = P(t, d)
* Price changes will be amplified by bundling supply and demand (e.g., through securitisation) so that P = A*P where A is either less than 1 (supply bundling) or greater than 1 (demand bundling) depending on whether buyers or sellers are being aggregated by a particular security.
* Transactions cost can be modeled as resistance (and Price * Transactions Cost will approximate Demand)
* External money supply can be modeled as capacitance (which will introduce a phase lag into price)
* Liquidity can be modeled as inductance (which also introduces a phase lag into the price)
* Demand can be modeled as voltage

For purposes of this model, I'm assuming that the external money supply obeys some predictable rules (like Taylor's Rules).  The system is going to be extremely indeterminate if the external money supply doesn't behave in predictable ways.  (There's a useful result right there!  Let's implement Taylor's Rules.)

From my very crude understanding of theory, this kind of model would permit the following predictions to be worked out from the nonlinear differential equations that govern such a system:

* Systems that include both inductance and capacitance (i.e., external money supplies and liquidity) are going to oscillate at a characteristic frequency.  That frequency is the "resonance peak," and it's amplitude will vary depending on the amplifiers.  If they're too strong, the circuit blows up.

* Systems that include large inductance but low capacitance (i.e., liquidity but no external money supply) are going to decay exponentially to zero price

* Systems that include mostly positive feedback (i.e., amplify demand without inverting or phase shifting the input signal) are going to increase exponentially (until they blow up).  (This was Monday's insight.)

* Systems that are tuned to include just the right amount of positive and negative feedback are going to oscillate stably within a limit cycle for long-periods of time.  In fact, such systems will "magically" self-correct price to demand.   Actually, this is most markets, most of the time.  We just haven't been paying enough attention to the bigger picture.

Somebody out there must have done some graduate school research on this topic.  We should send them to Bernanke and hope for the best.

UPDATE: Thanks Google.  Here's Steve Fairfax of MTechnology making a similar point.  And here's Donald D. Hester and D.L. Brito making a similar point... in 1974!  Do we ever learn?  (Don't answer that.)

UPDATE2: Here's a book on "Economic Dynamics" with a whole section working out a version of control theory applied to a more complicated model.

UPDATE3: I've worked out a numerical example with estimates for the subprime mortgage market here.

Modeling the Subprime Mortgage Market as a Closed Loop Oscillator with only Positive Feedback

Electrical engineers who design analog circuits learn an important rule of thumb: closed loop systems that provide only positive feedback to their inputs can become unstable.  In the diagram at right, when A and B are relatively good amplifiers, the multiplication of their amplifying effect at the input Vin can cause the system to blow up.

The closed loop oscillator with only positive feedback is actually also a useful way to understand what caused the subprime mortgage market bubble, and what could have been done to prevent it.

According to the newspaper accounts, something like this was happening: home buyers were bidding for price (Vin), which was being approved by assessors. Banks were then offering loans based on the assessments to the buyers, and selling the loans to loan-buyers.  Loan-buyers were then bundling the loans into CDOs (amplifier A), and then selling them to bigger loan-buyers, who were then bundling CDOs into CDO^2s (amplifier B).  The bundling permitted for larger-scale buyers of loans to get into the market, which has a multiplying effect on demand for loans, which in turn gave the banks and assessors an incentive to offer loans at higher assessed values regardless of the underlying value of the property.  There was no mechanism for negative feedback on price to enter the closed loop that buyers, assessors, banks, and institutional investors formed.  And when A*B approaches 1, as it will eventually in a system with powerful amplifiers and no negative feedback, the system blows up.

Electrical engineers avoid these kinds of instabilities (usually!) by building negative feedback into the loop -- i.e., by introducing some circuit element that inverts or phase shifts the signal fed back into the input of the circuit so that the signal doesn't just keep growing and growing until it blows up the circuit.

Lee Fennell at the University of Chicago law school has a great suggestion for how we could have built negative feedback into the system: bifurcate on-site and off-site risks associated with homeownership, and make buyers pay separately for these two types of risk.  So for example, a home buyer might be required to buy an insurance contract at the same time they purchase their home, the premium for which is large enough to permit the insurer to pay the home buyer off for a loss in home value should the value of her home later decrease precipitously because of off-site problems in her community (such as factory closings, pollution, or crime).

Now imagine how the same cycle would have worked: the home buyer makes a bid on both the home and the insurance for off-site risk -- the law has to require her to buy both, otherwise this doesn't work.  The home assessor approves an inflated price.  That's still the assessor's incentive because of the demand for mortgages.  But the insurer then raises insurance prices in response to the assessment to offset the larger risk associated with the inflated price.  If the home buyer can pay the increased premium, it's a deal.  If not, the deal falls through, and prices tend to fall back to a level that more accurately reflects the underlying value of the property.

UPDATE: Upon reflection, I realize that the government actually does not have to require homebuyers to buy both the home and the insurance.  In fact, only the insurance-provider and mortgage-provider would have to agree.  This raises other complications (what happens when they're owned by the same entity, e.g.?), but it's useful to see at least that new legislation might not be required.

Can the law firm business model be fixed?

The billable hour is a terrible metric for either the value of client services or of associate contributions thereto -- its two primary uses.  The trouble is that there seem to be few alternatives.  The reasons are obvious enough in each case.  Clients are good at knowing whether they're in legal trouble or not, but bad at knowing whether their outside counsel is working efficiently to keep them out of legal trouble.  Partners are better at knowing whether associates are working efficiently -- they know whether they're getting the work they need and also what they're billing the client for that work.  Unfortunately, partners have a perverse incentive to bill more rather than less for associate (and partner) time -- their own bottom line is tied to the billable hour.

To most stakeholders (client, partner, associate, employees), this looks like an train wreck.  In other parts of the economy, principal-agent problems of this scale could not persist.  But the lack of client transparency into the quality of legal services seems to be keeping the business model on life support.  Can the business model be fixed?

The practice of law is a diverse and varied pursuit.  There cannot be a solution that will work for every field.  To wit, there are some areas of law (such as plaintiff class action work) where alternative fee structures are the norm.  Thus, improvements to the business model will probably be area-of-law specific.

I'll focus on corporate transactions.  Here is my humble suggestion for how companies might get more efficient work from outside corporate counsel for much transactional work: bundle insurance and legal services.

Part (not all) of why corporate management pay for D&O/E&O insurance and transactional lawyers is for a similar economic purpose: to hedge against the downside risk of future lawsuits.  There are, of course, a host of other reasons why managers pay for good outside counsel, such as information about the market.  Outside counsel that specializes in a particular type of transaction may be far more knowledgeable about what is going on in a particularly industry than managers who may be entering for the first time.

Nonetheless, the downside risk of a lawsuit is something that good transactional lawyers will seek to avoid by drafting and refining transactional documents with an eye toward future plaintiffs.  And many corporations take a "belt-and-suspenders" approach by paying for both insurance and extra hours on the proxy statement.

Why not bundle D&O/E&O insurance with transactional legal fees in the following way: inform outside counsel that their pay will be based on an hourly fee that reflects the fair value of the information and services aside from the hedge against lawsuits, then pay a bonus at the end of the year based on the difference between what the insurance premiums are and what they would have been had the insurance policies paid out.

The benefit here is that insurance companies are good at quantifying the value of downside risk avoided.  By bundling the costs in the books, company management could make a more accurate assessment of that part of the value of their transactional services.