In Defense Of Bubbles

There is nothing the tech media and the broader press likes to ask me about more than bubbles.

“Is Snapchat a bubble?”

“Is Uber a bubble?”

“Is Facebook a bubble?”

“Is seed investing a bubble?”

“Is growth investing a bubble?”

And on and on and on.

It’s like bubbles are a disease that we need to eradicate.

Don’t get me wrong. Bubbles are something investors need to be careful with. You can make money on the way up but lose it all on the way down. I’ve done that. It hurts.

So at USV, we are careful to invest early in cycles and get defensive later in the cycle and take profits when they are available. If anything, I think we have been too conservative in this regard.

However, as I pointed out in a conversation with my colleagues yesterday, bubbles are a necessary part of any technology cycle, large or small.

Carlota Perez talks about this in her seminal book on technology cycles, Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages.

But maybe the best thing written on this topic is my friend Tom Evslin‘s blog post from January 2005, when the investment world was just waking up to the fact that the Internet bubble wasn’t the end of things, but just the beginning. I particularly love how he ends that post:

Historically, the results of bubbles have usually been more empowerment for more people.Β  Historically, bubbles have provided an explosion of funds which blasted away the entrenchments of an old oligarchy not only to the benefit of entrepreneurs but also to the benefit of consumers in general.Β  Think of the constantly falling price of transportation and communication.

If we should find a way to stop bubbles, if we were to put the genie of irrational exuberance back in the bottle, the winners will be whoever are the incumbents at the time and the losers will be all those who could benefit from another great breakthrough in infrastructure like railroads, canals and the Internet.

Bring on the next bubble.Β  And invest in it at your own risk. I will.

#stocks#VC & Technology

Comments (Archived):

  1. pointsnfigures

    There is no bubble in VC. As a percentage of the S&P, money in VC funds is about where it was in 2004. If anything, institutions are under invested in VC. VC has done better as an asset class than PE since 2006. We are seeing ground breaking inventions in every business sector daily that will make our lives better. It’s a good time to be writing checks in VC.

    1. William Mougayar

      You’re looking at it from an LP / asset allocation perspective.

      1. pointsnfigures

        I am. At the same time, there is no bubble in tech companies. It’s been largely pricked. The mom and pops that Taffer is referring to would be the dumb money that bought the IPO. Government regulation, and government monetary policy is causing what many see as a later cycle VC bubble. At early and Series A-C stage, I don’t see a bubble. I do see LOTS of startups compared to ten years ago. But, that isn’t necessarily a bubble. It’s people sensing the same opportunity.

        1. Lawrence Brass

          It is also capital and infrastructure availability, and markets with lower barriers of entry for the small entrepreneur, as never was before. Some of these bubbles are made of a new kind of soap.

    2. Taffer

      There is a bubble in tech companies. When you have mom and pop investors (albeit rich ones) investing in specially designed Morgan Stanley vehicles which then invest solely in a privately owned business (Uber) then you are in bubble territory.

    3. Tom Labus

      There always a “bubble” or period of insanity in the public markets. It’s a question of when to exit that is the challenge

      1. JLM

        .That is, by definition, a trade not an investment. No brief for it one way or another but there are some good trades out there.Some on the way up and some on the way down.Amazon turning solidly profitable? That’s a pleasant surprise.JLMwww.themusingsofthebigredca…

        1. Tom Labus

          I have a very hard time being short. The summer 08 trades, I could never make. There can also be a bubble in a longer run which compounds decisions for long term investors. Bezos may be one of the most ambitious CEOs we have seen

          1. JLM

            .I have been skeptical about Amazon but performance is performance. Going short is always a challenge.JLMwww.themusingsofthebigredca…

  2. Tom Evslin

    Thank you, Fred. You and USV have done a great job of funding the “inflationary” (in a good sense) phase of creatively destructive bubbles. Remember when calls to China cost a dollar a minute? And when 2400 baud (.024Mbps) was blazingly fast?

    1. panterosa,

      “Creatively destructive” – so the bubble’s patron saint is Shiva – creator, protector, destroyer.

      1. Tom Evslin

        yup. and Ganesh is the proud godfather

        1. panterosa,

          Exactly!! I have a ganesha at home and in my studio for this reason!

      2. Vasudev Ram

        It’s a bit different, actually:Creator: BrahmaPreserver: VishnuDestroyer: Shivahttps://en.wikipedia.org/wi…

        1. cavepainting

          True but Panterosa is also correct. Some believe that the ultimate godhead from which all other Gods and more generally, all forms originate is Shiva. For these believers, matter and form is just an illusion, everything is energy. And Shiva embodies all that is energy from which everything arises and subsides. What gets created gets destroyed and what is destroyed gets created again.Also, to note that the notion of Shiva precedes Hinduism or any religion. It is a state of nothingness that contains everything.

          1. panterosa,

            SO DEEPLY ZEN

          2. Vasudev Ram

            Yes, that is also called Brahman [1] and even other names. Hinduism is very complex and multi-layered and also contains within its umbrella term, many different schools of thought and belief, and not all of them date from the same century either (or maybe even millenium).Some consider that the Upanishads [2] distill the core ideas.[1] Edit: Brahman – not to be confused with either of Brahma (a god in the trinity linked above) or Brahmin (a caste) or even Brahmana (sections of some religious texts).https://en.wikipedia.org/wi…Not an expert on any of above though, just speaking as a lay Indian who has heard and read about some of this stuff.[2] https://en.wikipedia.org/wi

        2. Vendita Auto

          Make room for the Borg

          1. Vasudev Ram

            No one can beat Shiva. Have you heard of his third eye? Incinerates even a Borg, in a nanosecond.

      3. cavepainting

        The technical meaning of “Shiva” is that which is not. Which means that which is not made of anything that we know and understand. And yet that without which nothing that we know of can exist. An interesting paradox !

        1. panterosa,

          I see Shiva as the great recycler of energy and goods. And ideas

    2. JLM

      .The Internet and its speed has become an important element of life.I live in ATX and have 1gig service. I am contemplating a move to Savannah — my biggest actual concern? Shitty Internet service.I have become addicted to 1gig service.I used to spend a little time in Steamboat Springs. I had NO Internet service. None.Now, I have great service and I spend a lot more time there.Clearly, there is something wrong with me.JLMwww.themusingsofthebigredca…

  3. Taffer

    For anyone interested in the deep links between state investment, speculative growth and technological innovation I highly recommend Bill Janeway’s ‘Doing Capitalism in the Innovation Economy’. Fantastic book, to quote Janeway, on the ‘banality of bubbles’.

  4. guillaume

    beautiful quote thank you, there’s no way we would all be there without “the genie of irrational exuberance” acting throughout history

  5. jason wright

    The Blobble.

    1. Tom Labus

      Now on Amazon Prime

  6. ZekeV

    Before 1999 and “irrational exuberance” how many ordinary consumers were aware of financial bubbles as a phenomenon? Now everyone knows about cyclical greed and fear, and is looking for bubbles in every story. How much does this pervasive awareness and discussion of bubbles affect the cycle itself. Can we avoid the worst excesses of bubbles through education?It’s not a rhetorical question, I really don’t have a thesis. But I am concerned that all of the infrastructure that produced 2008 is still in place, and the printing press is running as fast is it can go. Not so concerned about government debt, but the magnitude and complexity of derivative products and hidden risks. Whenever I talk to a lawyer in securitization or other structured finance field, they tell me they are busier than they ever were including at peak of 2007. But I would have a hard time finding any particular asset type that is obviously mispriced.

    1. Cam MacRae

      magnitude and complexityBut I would have a hard time finding any particular asset type that is obviously mispriced….and that was when I ran for the hills, Your Honor.

    2. Girish Mehta

      Re | But I would have a hard time finding any particular asset type that is obviously mispriced.Because everything is mispriced today. In a world where everything is mispriced, it is difficult to find something “obviously” mispriced.The US 10-year yield at 1.55% does not seem so bad when the Swiss 50 year yield has gone negative (Swiss 2064 paper is negative).But, Comparing to the 1999-2000 period,Nov 1999 – Fed funds rate at 5.5%Jan 1 2000 – 10 year yield at 6.6%Today – Fed funds at 0.5% and 10 year at 1.55%.The 10-yr sounds not-so-unreasonable…when you look at where the rest of the world is and the Fed funds is.Truth is – Interest rates have lost their “signal” value. Everything is mispriced.

      1. JLM

        .From a guy who had a floating rate loan on a $100MM real estate deal on 19 Dec 1980 (Prime Rate hit 21.5%), this issue is really about the business cycle and the boom/bust momentum of that same business cycle.I never actually cried but I wanted to. When you live through something like that, you get some perspective. 21.5%! Prime + 0.25% interest rate.It is a bit different terminology made even more complex as it is not universal. It can be a bubble in one area and a bust in another area. It is further exacerbated by the reality that really smart folks can disagree as to what is really happening.Uber was in deep bubble territory last month but it is something else, perhaps, after it merges with Didi in China which then puts them in partnership in the US with Lyft by virtue of Didi’s investment in Lyft.Cipher that one out.It is also a great example of how an entrepreneur can make a dramatic impact on his company’s future by taking a bold step.I don’t think anyone saw this coming but when one realizes Uber had pissed away $1B each of the last two years to go mano a mano in China with the hometown team, you realize that eating a little humble pie can be a very important thing.Think about it — a billion dollar a year cash requirement evaporated overnight because Travis Kalanick had to admit he was getting his ass handed to him. That is the most positive thing I’ve ever learned about the guy.BTW, the US prime (which isn’t as relevant as it once was) had been 3.25% for seven years — evidence of how shitty our economy is and how we’ve entered the Twilight Zone of the Japanese Third Missing Decade.Like they say in the Oil Patch, “Dear God, just one more boom and I promise not to piss this one away.”JLMwww.themusingsofthebigredca…

        1. LE

          Uber had pissed away $1B each of the last two years to go mano a mano in ChinaNo nexus no washie.The fix is in for the local company. Uber bought local knowledgeable talent that additionally must have been connected in the right ways. But that was not enough.It’s like Trump says. They cut people’s heads off and we can’t even waterboard. As Americans we bend over and grease up like the honorable guys that went down with the Titanic.

          1. JLM

            .Trump is right about the nature of our trade disadvantage in introducing American products and services into markets which export to ours.We need free, fair, Trump trade.http://themusingsofthebigre…JLMwww.themusingsofthebigredca…

          2. LE

            The average politicians intuitively has no seat of the pants feel for leverage, deal making and quid pro quo when it comes to business. In politics yes. In business no.

        2. Girish Mehta

          Didi are now reported to be investing further in GrabTaxi who will use these funds to step up the fight against Uber in SEAsia.http://qz.com/749115/days-a

          1. JLM

            .Danegeld, Kiplinghttp://www.poetryloverspage…Once everyone figured out how to call and respond, now it’s back to being a very simple business. It is not rocket science.JLMwww.themusingsofthebigredca…

          2. Girish Mehta

            Had not read that one…Thanks.

          3. sigmaalgebra

            Ah, who are you going to believe, that old guy Kipling or our very own POTUS Obama! πŸ™‚

        3. Girish Mehta

          The Bank of England was established in 1694. Today it cut interest rates to the lowest level since 1694.The most surprising thing is how totally unsurprising today’s move was.Source: Bank of England and BBC

          1. Lawrence Brass

            It maybe has to do with them just doing the stupidest thing since 1694.Watch it Trumpians.

        4. sigmaalgebra

          There is a claim that in China Uber was trying traditional predatory marketing practices, in selected areas, under selling the competition, and losing money doing so, to drive the the competition out of business. Then, the Chinese Communists, not all totally stupid, laughed at Uber, called their move “cute”, that is, silly, trivial, childish, hopeless, claimed it violated Chinese laws or some such, and told Uber what they had to do in China or just leave. Or, in US terms, Didi made Uber an offer they couldn’t refuse.Broadly, long term, for anything like their current business model, I don’t see that Uber has any chance at all anywhere. Why? Uber has next to nothing for real barriers to entry or significant market advantages. And, it’s just too darned easy for locals, city by city around the world, to compete with them, and the locals will always have lots of local advantages, e.g., in City Hall.

  7. cavepainting

    What seems like a bubble is one until it is not.What seems like not a bubble is not one until it is.Nothing reflects “non-consensus and right” as the metaphor of bubble and the minority who call it against the popular grain of thought.You make money by betting against and for bubbles.

  8. johnboehmer

    Bursting, or hissing, bubbles are always substantial learning opportunities. Take advantage of them.

  9. James Ferguson @kWIQly

    The first stage of inflating a balloon is hard, as the surface tension per unit increase in volume is considerable.As inflation occurs, the rate of growth must fall, but each incremental investment is easier (but less proportionally to the whole) the walls of the bubble are thinner and less resistant to growth (or penetration).When value perception relies on the rate of growth – (like a Ponzi scheme) all bubbles are doomed.Otherwise, If value can be matched to a proposition that sustains money in vs money out the balloon can be tied off. If not the skin gets weaker and it bursts (catastrophically, with a noise or quietly with a whimper and wheeze).The most sustainable Bubbles relate to a process that becomes a beneficial habit, rather than a problem cure (which deflates future potential).A sustained bubble seen from a distance is merely an incremental market change.All change starts with a bubble!

  10. TeddyBeingTeddy

    Fred what does your diligence process typically look like? How thorough/deep do you go on the biz vs the people?

    1. fredwilson

      it all depends on facts and circumstances. there is no normal for us on this.

      1. TeddyBeingTeddy

        Makes sense

  11. kidmercury

    more often than not, bubbles are a result of monetary policy that transfers wealth from lower classes to higher ones through artificial money creation/low interest rates. as such, bubbles contribute to economic inequality in addition to capital misallocation.while bubbles may occur without monetary malfeasance, they are neither necessary nor inevitable. when bubbles are celebrated, the poverty they directly create should be noted. there is a reason that bubbles in silicon valley result in rising real estate prices in the area and animosity towards technology companies. likewise, the housing bubble has contributed significantly to the difficulty the millennial generation (and perhaps subsequent generations) has with home ownership, which is at multi-decade lows in the US.

    1. Tom Labus

      South Seas, Tulips??? Maybe more a part of human nature

      1. kidmercury

        that is probably the best example of the point i’m referring to. money supply in 1630s holland exploded, in turn leading to the tulip bubble.as noted this is not the only cause of bubbles, but more often than not it is the primary cause — and its results are not innocuous.

        1. Tom Labus

          Interesting call. Are you sourcing some book?

          1. Cam MacRae

            https://link.springer.com/a…If you can’t access the full paper and have a couple of hours to burn feel free to hit me up on first.lastname at google’s mail service.

          2. Tom Labus

            Thanks. Looks good

          3. kidmercury

            the captivatingly titled “Early Speculative Bubbles and Increases in the Money Supply” https://books.google.com/bo

          4. Tom Labus

            Thanks.

        2. LE

          money supply in 1630s holland exploded, in turn leading to the tulip bubble.I don’t know if I agree with that at all. (Neither do many economists from what I can tell). While monetary policy might have made it more possible (or helped) it doesn’t explain why in this case tulip bulbs were the desired object. Why not something else? Why not inflation in multiple objects? (Maybe there was but not to this extent you will agree).Otoh in housing there is little dispute (common sense) that the availability of mortgage money led to the quick rise in housing prices. But also in that the greater fool mentality was in play. But the primary driver was the availability of money.

    2. SubstrateUndertow

      Aren’t we conflating two distinct categories of bubbles here ?Monetary policy bubbles vs technological disruption gold rush bubbles.

      1. kidmercury

        the most generous statement i can make to the perez ideology is that the 2000 bubble was both a monetary policy bubble and a “legitimate” bubble, and thus its unbelievable heights. but in general, monetary policy is the key independent variable.

    3. Rob Larson

      How do low interest rates = a transfer of wealth from the poor to the rich?I’m afraid I missed a key step or two in your explanation. Thanks.

      1. kidmercury

        sure. here is what is happening:1. first, it needs to be understood that all money comes from debt; money is *created* when debt is issued. this may seem counterintuitive but it is a fact. 2. so, when interest rates are low, more debt is taken out since debt is cheap. this means more money is created. so, artificially low interest rates tend to increase the money supply significantly. 3. this ultimately ends up resulting in price inflation. this inflation is not uniformly distributed and thus certain sectors may experience more price inflation than others — and thus, we see bubbles — prices in general will rise. 4. inflation is a regressive tax, because it increases the price of assets AND consumption. poor people don’t have assets; richer people have them. so inflation causes everyone to spend more on food, housing, cars, stocks, etc. if you own housing and stocks, this may not be a big deal and could even be advantageous. but if you don’t own a house or stocks and you still need to pay for consumption, you are still paying higher prices.many people believe inflation is not occurring and that thus this argument is not true. rather than government’s inflation numbers, it is better to focus on things one actually has to pay for to survive to see where the inflation tax is appearing. housing, education, automobiles, tolls, etc. the chapwood index is a useful statistical measure in this regard: http://www.chapwoodindex.com/

        1. Rob Larson

          Thanks – that’s a thought provoking argument. Couple comments:1- do you have data for #2 above? I agree that borrowers have an incentive to borrow when rates are low, but on the other side, banks have a disincentive. Banks love to lend when rates are high. And when banks are nervous AND rates are low, they really hold onto their money. Recall all the articles about how frustrated the Fed has been since 2009, trying to convince banks to lend more, but the banks weren’t having it.Also, the rates might not be that important of a driver for borrowing anyway. Businesses borrow when they see opportunities to invest in which are significantly better than the borrowing costs. During times when they see lots of opportunities and have a lot of confidence about those opportunities, they tend to borrow a lot, regardless of the rate environment. When they are fearful, they stop borrowing, regardless of how low the rates are. People are similar – people have on net been paying down their debts over the past 6 years, despite record low borrowing costs, because they are still skittish about the recent recession.2. the best argument that inflation is bad is that it hurts older people who are retired and have saved up for their retirement – making the purchasing power of their savings less than they had counted on when getting ready for retirement.Your argument that it makes assets and consumption more expensive seems more geared toward poor people still working. That argument could be true IF wages do not rise with inflation (otherwise it’s a wash). My understanding however is that wages do in fact tend to rise with inflation. For example, see https://www.stlouisfed.org/… If your data shows the contrary I’d be interested in seeing it.Thanks

          1. kidmercury

            the easiest way to see it is to look at a chart of money supply — http://www.economicgreenfie… — and think of interest rate policy. you’ll notice that the steepest inclines in money supply correspond to the stretches of low interest rates.there are many criticisms of how CPI is calculated and whether it is accurately reflecting what people spend money on, including things like health insurance, tolls, debt payments, and college education among others. there are some alternate CPI measures that most proponents of the argument i am putting forth here would utilize instead: http://www.shadowstats.com/

    4. Roboticus Aquarius

      It seems almost axiomatic that loose monetary policy will tend to enable bubbles, but I expect the bubble phenomenon is as much a result of new technology and ‘animal spirits’. I’m not so sure the results are all bad. Arguably, too many people were buying homes. Was it appropriate to ‘reset’ the market, to encourage a return to smaller homes and more restrictive financing requirements? (Bubbles are only one factor, so I think the impacts of globalization have taken that potential option out of play, but I think the question is still legitimate.) I know a lot of people who as a result of the internet bubble had new high-paying jobs with fancy titles, etc, etc. Many of those people eventually lost those jobs, but in the meantime they made very good money (which they saved, if they were wise), and got experience of the kind that is very hard to obtain in a less frothy economic environment. Or Wyoming in the recent oil boom – 2 years later, it’s a bust, but oil workers who were willing to move made far more money than is generally available in that field. It’s difficult to respond to quick changes; there will always be those who benefit and those who get hurt from any change – I’m just doubtful that the results are negative on balance.

      1. kidmercury

        i don’t say the results are all bad, only that the results contribute considerably to poverty and income inequality.your anecdotes refer to beneficiaries of the bubble, who of course will have happy stories to tell. nationwide, though, the statistics are clear. housing expense as a percentage of income continues to grow.http://www.bloomberg.com/ne

        1. Roboticus Aquarius

          I think the comment about beneficiaries of the bubble is a red flag – during 1996-1999 who didn’t benefit? I think you have to look at very isolated instances (blue collar workers in silicon valley, for instance) to make such a case. On the other hand, I am very aware that the fallout hurt many people, many of whom still struggle, but the evidence linking that to bubbles seems quite weak. Take the Bloomberg article you link. Is the lack of housing affordability an artifact of a bubble, or is it a result of globalization? It’s far from clear from any comment made by the author, nor do they explore other possible causes. Anyways, I see globalization as by far the more damaging force to the US, and how do you segregate that from the fallout of the internet and derivative crashes? And, let’s accept the contention that more people are paying a higher percentage of their income in general… so many other things are now a smaller percent of our income. Virtually any technology purchase is far cheaper. A new base-line Camry costs about what I paid for the one I bought 20 years ago (300K miles, wahoo!). There are always people with relatively lesser wealth and income. I just don’t see the linkage to bubbles, per se.I think the connection of the money supply to bubble behavior is information that’s good to confirm, I’m just not there on the rest of it.

          1. kidmercury

            as for who didn’t benefit from 1996-1999, the same concept applies even then. see the chart below, and note when the huge spike began.the co-occurrence of rising housing costs and bubbles is not coincidence. both share the same cause, which is inflation of the money supply, which results in more money chasing the same amount of goods — and thus higher prices and bubbles.http://4.bp.blogspot.com/-k…as for your camry, no, new car prices are rising (because of inflation of the money supply). plus you have now 7 year car loans, and a car loan market that is on its way to being the next subprime….http://z822j1x8tde3wuovlgo7

          2. Roboticus Aquarius

            And yet I believe the money supply was pretty flat all the way until 2008, M0, M4 whatever you want to use. That doesn’t mesh well with the home price chart you show above. I don’t think anyone would disagree that when the money supply increases, assets will increase, but that’s a long way from showing that a bubble, per se, is net of all impacts, bad for the economy… only that there will be some who benefit and some who are hurt, but that’s true regardless of whether a bubble exists. There are a lot of assumptions in the linkage you propose that seem far from proven. Granted, this is just a comments section, but the side bars I brought up are still legit. A new Camry is only about 10% more than 20 years ago, and car prices on average actually declined last decade, only resuming a modest climb fairly recently. That increase is far, far less than general inflation. If you want to mention car loan length as proof of increasing price or cost, then you need to show the related information such as, is that an attempt to expand the car buying population? I agree these financing innovations are dangerous, but you imply it’s due to lower salaries in general, and that may not be true at all. It’s an interesting question whether bubbles are truly a negative for the general economy – certainly volatility makes life difficult for many – and I am sure there are some negative impacts – but as a generic statement, I don’t think you can hang your hat on it yet.

          3. kidmercury

            Sure, there are many numbers out there and so people can choose to believe whatever numbers they’d like. Here is a chart of m2. Note when the large spike begins…..https://www.google.com/sear

          4. Roboticus Aquarius

            I would think it’s more a case of recognizing which numbers are relevant, and when it comes to money supply, they probably all are to one degree or another. It looks like M2 is the only measure that supports the contention that money supply drove housing, and that only modestly – 1996 shows a slight increase in the growth rate (after a mild decrease), with no real spike at all until a small one (comparatively) in 2008, as I mentioned above. Interesting, as this would imply the housing situation in the late 1990’s may have been more an industry issue, with loose housing loans cycling into derivatives, rather than an economy-wide phenomenon. In any case, it’s always tougher to argue in support of an assertion than against, appreciate your willingness to discuss.

  12. ZekeV

    There’s a lot of silly money in bitcoin-related projects. I don’t really see it as a bubble yet, because bitcoin is still perceived (rightfully) as risky, and the big money has not started to flow. Banks, prop traders, and even venture investors are still dabbling. On the consumer side we are still waiting for the first S-1 to get approved for a bitcoin-backed ETF (to say nothing of a litecoin-backed ETF, etc.). If all of bitcoin went away, there would be some sad puppies in NY and SF but no systemic impact on the national economy. That will change…

    1. William Mougayar

      I agree, but there are bubble formation signs in the works. Maybe with the blockchain, given it’s decentralized, it might manifest itself as a series of small bubbles. Eg. the DAO fiasco was a mini bubble of sorts.

      1. ZekeV

        Yes, and it is a bit scary to lurk in the altcoin forums. I just know that people who absorb the prevailing naive noise about bitcoin are going to lose their shirts. I try to avoid reading the sob stories from people who lost everything in The DAO or in failed exchanges. Too painful!

  13. jason wright

    Negative interest rates.

  14. Anny Smile

    Thanks for the article! Very interesting and informative one!

  15. Sam

    Sounds like the argument is: the late 90s bubble enabled the massive internet infrastructure investment that was required to create scale efficiencies and network effects underpinning the current age of value-creating startups.Bubble -> Infrastructure Buildout -> Value CreationRaises two questions:Looking back… Could you have gotten that infrastructure buildout without a financial bubble?Looking forward… If we are observing a financial bubble inflating now, is it leading to the creation of a useful building block for the future? Or is it just tulips?

  16. sigmaalgebra

    There is nothing the tech media and the broader press likes to ask me about more than bubbles.Darned right the do: They know that bubbles are commonly too often when they burst not just destructive but actually very DANGEROUS.Okay BubblesSilicon Valley bubbles, or as Doerr might say, not a bubble but a “boom”, fine.Not Okay BubblesBut the bubbles we need to avoid were in 1929, likely the one in 1999, and definitely the one in 2008.A Really Bad BubbleThe 1929 bubble led to the US Great Depression. Spreading through world trade, that led to Tojo and Hitler and WWII and killed around the world ballpark 100 million people.Just the 1930s caused losses of psychological capital in the US that we still have not recovered from. Too much of the 1930s in the US were with conditions humans can’t put up with without serious damage or death. E.g., a lot of women and girls learned that husbands and fathers could not be counted on to take care of them. This was mostly not true in 1800; it was definitely true in 1930.How to Get Out of a Bad BubbleWe didn’t get out of the Great Depression until people started shooting at us, and then, with essentially a command economy, we got out in about 90 days flat at which time essentially everyone had multiple job offers and often 2-3 jobs. And nicely enough, when the shooting was over, we stayed out of the Great Depression — a lot of people were very unsure about that.Socialist EuropeWhy is Europe so interested in socialism, the EU, and the Euro? Sure: They remember WWII. There’s no way they could forget WWII, not even for another 100 years. Same for England, Poland, the Baltic states, Norway. Russia can’t forget. I don’t understand Asian cultures at all well, but I doubt that the Philippines, Viet Nam, Korea, Japan, or China can forget.More Recent Bad BubblesFor the bubbles of 1999 and 2008, Trump is correct: In real terms, a lot of people in the US, if they do still have a job, are now making no more in real terms than they were making in 2000. Really big, bad bummer.Chair Bernanke saw the threat of the 2008 crash when he said “I don’t want to be the Chair of the Fed that presided over the second US Great Depression.” or some such. The threat was real. The good news is that he did keep us out of a second great depression. The bad news is that we got a Great Recession, now 8 years long, 2/3rds as long as the Great Depression.As above, we know what got us out of the Great Depression. Apparently so far we don’t have even as much as a weak little hollow hint of a tiny clue how the heck to get out of the Great Recession. We’ve used our low interest rate card and our fiscal stimulus card, and we still are not growing our way out. We know one way to get out in 90 days, but apparently we don’t want to use that one and can’t find another one.Big Bubbles Are BadBig bubbles are really bad; very broadly they are wildly, enormously, hideously destructive: As prices rise, people who saved cash are ripped off big time. As the bubble bursts, there is huge destruction and then years of wasteful idleness.The Good Old Days without BubblesIn the US economy of, say, 1800, f’get about bubbles: People lived on farms and were nearly 100% self sufficient within a radius of about 20 miles; that is, what happened to the economy more than 20 miles away, say, in the cities, didn’t much matter. But now, what happens in the cities is 90+% of the economy and very much matters.Economists and BubblesSooooo, our macro economists need to find ways to make our big city economy as stable as the farm economy of, say, 1800.Okay, macro economists, you’ve got a lot of work to do. And they haven’t done the work and are not even close to getting it done.In our city economies, we just flatly don’t know what we are doing. We are standing on something we have built that is rotten and tends to collapse frequently for meager reasons and hurt/kill millions of people. The next collapse could trigger WWIII and kill nearly everyone.First CutGeneral, first cut approach to fixing a network of flows: Look at the bottlenecks and alleviate those.For a first cut, look at the economy, say, look at it now, and see where the bottlenecks are. Then attack those. Are the bottlenecks fiscal or monetary? Nope.Okay, then, find out where the heck the bottlenecks are. Sarbanes-Oxley? Too restrictive regulations on commercial banking? Too much in imported goods? Hysterical, earth-worship EPA? Too much outsourcing to other countries? Too much in US profits parked overseas? Too much in various business regulations? Energy prices too high? I don’t know. But, first cut, we need to, and darned well better, find out.This is not the first time I saw that the economists were for the economy much like the witch doctors with boiled rat tails were for medicine. As a grad student, I considered being an economist. But just a first cut look at economics convinced me that the whole field was, right, like grotesquely incompetent witch doctors and so bad there was no hope for the field or an opportunity for me.We better fix the witch doctor grotesque incompetence of macro economics. A fix is worth trillions, maybe all of human life on this planet. Biggie problem. Bigger importance for a good solution.Now if all the economists were lined up end to end, it would be a good thing.SummaryBubbles? I’m not really happy about bubbles.

    1. SubstrateUndertow

      To be fair economist are required to model an impossibly complex system of interdependencies between 7 billion volitionally volatile/panicky agents all running behavioural algorithms/memes centred on short sighted localized self-interest.Pile on top of that mountainous layers of bureaucratic institutional agents all operating on the same basis.Economist are largely using linear tools to model the organismic dynamics that underpin human economic interdependencies with no widely available/appropriate organic big-picture framing/language-set at hand.Even if god were to hand delivered us a clear/effective model of all those organic interdependency-bottleneck and the economic mechanisms by which to solve them we would be hard pressed to implement them due to the tenacious evolutionary inbred power of volitional-selfinterest all running short-term/localized interference !We may have reached our evolutionary hull-speed ?Economics is now a global organism of networked dependencies where, as always, the network’s global success/coherence/stability/sustainability pivot on effectively synchronized signalling/language behaviours.The penny in the currency for solving those global organic interdependency-bottlenecks is a set of shared language/memes that allow us to effective visualize/debating all these organic economic dynamics/dependencies.It is the power of abstract shared language effectively mapped onto our presently pertinent survival challenges that truly separates men from bears.Language MEDIATES everything in human social evolution/survivalOR if you willthe medium is the message Where are all our organic language explorers !

      1. LE

        To be fair economist are required to model an impossibly complex system of interdependencies between 7 billion volitionally volatile/panicky agents all running behavioural algorithms/memes centred on short sighted localized self-interest.Equates to “roughly less accurate than weather forecasting..”

        1. SubstrateUndertow

          The Economy Isn’t A Machine. It’s Organic and Constantly Evolving.

      2. sigmaalgebra

        > To be fair economistOkay, for a minute before we lay them end to end.”Linear”? Yup.A lot of that is well put. I agree with nearly all of it. There’s a point about the necessary important role of language.BUT: Gee, astronomers looked at the tracks of the planets — they wandered and looped. But, Newton nailed it. Shoot in some alpha particles and notice that there are some hard points in there — yup, understand atoms. Atoms have layers of clouds of electrons interacting, and then molecules are much worse. But eventually chemistry straightened it out. Biology? Inheritance. Yup. How the heck? Ah, X-ray diffraction. Structural chemistry. DNA! Sequence the stuff? Chop it up; sort it; sequence the pieces; use a computer to put the pieces back together!Uh, short story: There’s a lot of really good, brilliant, productive, crown jewels of civilization stuff done. But not in economics.

        1. Lawrence Brass

          Having trouble decoding @SubstrateUndertow:disqus organic language? I am a fan and have taken some notes that might help. :)I wonder if we have enough computational power and a matching methodology or math or language to model really complex dynamic systems. My guess is that we have all the pieces but we still don’t know how to connect them or maybe it is that the complexity of the problem exceeds our individual comprehension.The only way out as Substrate implies (i guess) is to connect the collective mind in new and different ways and at higher levels, as Pokemon hunters do when they hunt collectively. This may sound ironic but it is not. If we organise collectively with a common goal and proper incentives and computing power we may solve together problems and reach agreements that otherwise won’t be solved or agreed. The network fabric is in place and that is where the pressure that inflates most current bubbles come from.

          1. sigmaalgebra

            I suspect that in understanding and predicting the economy, at best we are where the astronomers were right before Newton.Maybe we could do a big simulation, simulate each person, company, product, service. Alas, that might too soon take us into supply curves, demand curves, propensities where we would be using mostly just survey data to make estimates. Then, even if that appeared to start to work, as the economy started to change as we want, we might be very short on good input data for our simulation.Simulation is such a simple idea that I have to believe that there people who have considered that and could lay out some serious challenges.For an optimization from any of that old operations research stuff, often applied to economic theory, sure, the computing could be absurdly large — e.g., stochastic optimal control as inE. B. Dynkin and A. A. Yushkevich, Controlled Markov Processes.The Soviets had no shortage of good mathematicians who wanted to apply such things to macro economic planning.Or, controlled Markov processes is in part what would love to do if had a lot of data have no chance of having. In the US, Samuelson considered such things.But for a simulation that was not much more than just basic budgeting, the computing should be reasonable enough.I don’t want to suggest that there is a good, easy solution, although I suspect that easy solutions are about all that will be tried.I can’t solve the US economics problems: I concluded that in grad school, and, now for some somewhat different reasons, I have the same conclusions.

          2. Lawrence Brass

            I was thinking about something more contemporary, in the lines of what that secretive algorithmic trading companies are doing but perhaps with a broader scope. Kind of Markov processes with live input feeds, if that have some sense, as that would make them non-Markov. (?)A big simulation driven by part live (and real) part random or modelled variables but sensible to the real inputs. I think that there is plenty of public data available, maybe not on the domains you need.Found this, somewhat related:http://www.financialsense.c…If everything fails, declaring the economy as an unstable and unpredictable system may be a thing to do for now.

          3. sigmaalgebra

            MarkovConsider a stochastic process, that is, intuitively, something, say, a process, varying randomly over time. Actually, the process is also permitted to be without any randomness, that is, deterministic. So, really, we are considering darned near anything that is parameterized (stay intuitive here, which is good enough, and avoids a lot of math notation impossible to type clearly into Disqus) by time.So, the process being considered can be (A) trajectory of a basketball, (B) flight of an airplane, (C) a company from startup to IPO and selling out via M&A, (D) sure, the stock market, an individual stock, a set of stocks, or the whole market, (E) arrivals at a Web site, (F) darned near anything that is parameterized by time. The time can be discrete or continuous.Well, still being intuitive, and from 50,000 feet up, the Markov assumption is a nice compromise — it asks less than some other setups but still has enough to do some work. Also what it does ask often intuitively seems okay in many real cases.So, the Markov assumption is that, for the process, at each point in time, the past and the future of the process are conditionally independent given the present of the process. So, right, the definition is in terms of conditioning in probability. The grown up version of this is from the Radon-Nikodym theorem with a famous proof by von Neumann.So, in particular, if we want to predict the future of the process and we have the present, then we have all we can use because the past of the process will do us no more good. If we don’t have the present, then the past might do us a lot of good. But, again, if we DO have the present, then the past doesn’t provide us any additional help. Process State So, then, intuitively, the present of the process can be regarded as the state, that is, in the context, all we can have to characterize the process.Well, for this, the state, that is, the present, has to have enough information to characterize the process. E.g., if temperature is crucial for predicting the future and if temperature is observable in both the past and the future, then likely to be Markov the present, the state, should also include temperature.Really, intuitively, the more information in the present, the more likely it can serve as the state and let us meet the Markov assumption.Well, sure, there is a special case: Let the state be the present and all the past. That should be enough information?But with the Markov assumption, it is hoped that commonly in practice the state that we observe in the present will be relatively small.Okay, for the US economy, what might be the state we observe in the present to meet the Markov assumption? Sure: Darned near every bank account, inventory record, etc. in the country. So, right, for the US economy, for meeting the Markov assumption, our state stands to be enormous. Not promising for getting utility out of the Markov assumption. Controlled Markov Processes Well, given a Markov process, maybe we get to apply controls, e.g., what a pilot does in an airplane. Then, suppose we want to know what controls to apply to get the airplane from, say, NYC to Chicago in least (averge) time. That is, we want a case of stochastic optimal control.Well, there is some nice math for that, and it is in Dynkin’s book. In probability, stochastic processes, and statistics, Dynkin is a world class guy, student of Kolmogorov and Gel’fand and long at Cornell.Then, sure, in principle could apply Dynkin’s optimization for Markov processes to getting the US economy back to where we want it in, say, minimum time. But, there are two biggie problems (1) getting the data and (2) doing the calculations. Still, Dynkin is such a good mathematician he is able to get some results just from the overall setup and without the terabytes of data and rooms full of super computers.IIRC, the current Fed Chair Yellen mentioned that maybe the Fed could use some control theory. Maybe she meant just deterministic optimal control theory!A lot is known: E.g., there is an LQG process — linear, quadratic, and Gaussian. So, the process is based on a linear system; the quantity we want to optimize is a quadratic expression in the state; and the source of the randomness is just a Gaussian process. In this special case, can f’get about the randomness, find the optimal control as if had a deterministic process, apply that control in the random case, and still have the optimal control. Cute.I tend to like that field: I wrote my Ph.D. dissertation in it. There for the problem I attacked, I was able to get the data and do the computing.You now have a nutshell version of Markov processes, controlled Markov processes, stochastic optimal control, etc.Again, however, I’m not going to try to be an economist. The first main reason is, I have a darned good startup I am pursuing. But a second main reason is, I’d have to work with economists, and I wouldn’t like or respect them; they would hate me; and I would soon come to hate them.

          4. Lawrence Brass

            Amazing Sigma, thanks for staying intuitive instead of using math notation, or should I thank disqus. πŸ™‚ To be honest, I sucked at college math. I have an inclination and a want to deeply understand things, sometimes I can be very obsessive about it, it is a bless and a curse because I am a bit slow.At engineering college succeeding in math courses consisted in solving the problems fast, no problem if you didn’t fully understand what you were doing. I suffered with that. On the contrary, everything that was intuitive and observable as applied physics was easier for me. Then I fell in love with machines and coding and that quite decided my future.So thanks again for staying intuitive and for the nutshell, very interesting indeed, shows that you have taught before. It is a pleasure to read you.Coincidentally, I have been working these days in reducing large data sets from router logs from an ISP. Some parsing, accumulating some fields for a given key, some minimum and maximum values. They were doing this with the tools that everyone uses because everyone else uses them, abstractions built upon abstractions, so performance was rather poor. An hour of generated data took an hour to process. This was a pre-sale showoff so they couldn’t solve it throwing hardware at the problem as most people do. The raw log files for a day summed up 17 terabytes.So they asked me if I could do something about it. Of course they continued processing in parallel to cover for the case I would take too long to solve the problem. At the end, after 8 iterations we end up with a C program processing roughly 800.000 records per second per thread, CPU bounded. Nothing really impressive for me because I am used to native and somewhat minimalistic development, as we have chatted before. But they were quite impressed. I asked them if they would be interested in an optimized version in the case they get the contract, the reply was that it was “good enough” as it was because the definitive process was described in the proposal as being built with Hadoop and other standard tooling. No interest in a multithreaded sharding version of the particular C solution.It sometimes feels a bit lonely when you don’t go with the herd.As for economists, I would care about being viable for them, not sure if their love is a requirement. Your customers should love you though.I have some intuitive doubts about the value of past data in a process. If you have data at the present, such as frequency or speed that have an intrinsic time component, I would say that you can’t compute that variable without looking at the past.Have a nice weekend with a lot of exogenous interruptions.

          5. Michael Pavey

            Interesting. In the Markov temperature example, why would the past not provide additional information? For example, if you are measuring the temperature of a glass of boiling water placed in a room-temperature environment over time, you would see a steady decline in the temperature. In the current state, you might see a temperature of 50ΒΊC. Wouldn’t it be reasonable to guess that the temperature in the future would continue to decline β€” something you couldn’t infer without seeing the past temperature data?

          6. sigmaalgebra

            Right. Soooooo, you just observed that in that case, don’t have a Markov process! Why? The state doesn’t have enough information. Soooo, maybe will have a Markov process if include the recent history of temperature.

        2. Roboticus Aquarius

          I don’t really agree. Newton was almost right, but his equations are very slightly imprecise, and he didn’t understand why his equations (almost) worked. There is still a ton we don’t understand about molecules, let alone atoms. Biology – how to cure the common cold? Cancer? Help those with Autism? Economics suffers from the impossibility of isolating individual factors. There is a lot of work done in economics that depends on highly suspect assumptions to support the wonderfully creative math used to identify relationships between risk and reward, employment and inflation, etc etc etc., but regardless, you end up having to work with incredibly unstable underpinnings.

  17. Richard

    If VC bubbles destroy wealth and wealth is concentrated to the LPs (the 1%), and it’s only an accounting number, no real harm here. Just keep it out of the public markets as in 2001 when VC was more of a ponzu scheme (though through great marketing it is never discussed this way).

    1. Realist50

      The LP’s are to a large degree public pensions, university endowments, etc. So ultimately somebody – often taxpayers – is on the hook if LP’s don’t get their returns.

  18. kevando

    The question “Are in in a bubble?” is tricky because it gets asked in the present tense, but it’s actually future tense. “Are the increasing prices of things going to start decreasing extremely rapidly soon?”Of course people are gonna ask an investor this question.

  19. JLM

    .”It kicks as hard as it shoots.”Here is a perfect example of how we allow ourselves to be worked into a lather about something that on the surface is pure baloney. Do the math — 130 hours per week? Really?http://www.businessinsider….The results, given recent events, do not speak well for either the bubble or the effort.JLMwww.themusingsofthebigredca…

    1. Cam MacRae

      I used to do overnights when I was a pup. I stopped. The chairman of our company — significantly more badass than Mayer — took me aside and told me he was singularly unimpressed. “You know who impresses me?”, he asked in his accent 30 years removed from St. Louis, “The guy who comes in and does his job smart and well, and goes home at the end of the day. The guy who takes a week of vacation every quarter”. He continued, “I have no time for people who light fires so they can put them out”.Consider me schooled.There is a word for people who crap on about how they work 130 hours a week and it rhymes with banker.

      1. JLM

        .For years I had a very productive partnership with a bunch of Englishmen I’d met in another endeavor. The chairman used to say, “We English and you Americans are separated by a common language.”The English got first pick on a lot of words — wanker being one of them.When I was getting out of the Army (MBA in hand), I had an interview at a silk stocking IB shop. They were keen for combat arms army officers (engineers with MBAs) and were using this “stress” interview technique.They posited the question about working late and asked me my reaction — to which I replied I’d help them get their work organized so there was no necessity to work that late.The young guy — younger than me — told me I didn’t understand the nature of the stress of IBing. I told him what I’d been doing for the last 5 years suggesting anything that didn’t entail incoming mortar fire was really a bit of a lark.Needless to say, I did NOT get the job though the older partner called me six months later and asked if I was still looking.It is hard to fathom the destruction of value that is Yahoo. One good investment and the rest a cess pool.JLMwww.themusingsofthebigredca…

  20. Supratim Dasgupta

    Anything REAL in the world has its UPS and DOWNS. Bubbles are DOWNS and without them VC would not be REAL

  21. Daniel Li

    Great article – economic cycles (and bubbles as a result) are an important part of a healthy economy that is able to make investments and loans to finance growth. Check out Ray Dalio’s great video on “the Economic Machine” for a look at why https://www.youtube.com/wat

  22. Adam Sher

    You can answer the bubble question with, “The answer is 42. Perhaps you’re asking the wrong question.”

  23. jbcolme

    I could not help it but think of Italy as an example of what happens when there are NO bubbles, or at least when is difficult for them to be. If there are no bubbles, everything is still. Which is just the brink before decline or collapse. Bubbles may be dangerous, but they can be a huge source of growth and innovation. No bubble at the other hand equals death.

  24. Deborahrdamron1

    <<l:y<my b=” room-mate Ammber Franklin Is getting paid on the internet $98/h”….!!bx589a:two days ago Black McLaren P.1 I bought after earning 18,512 Dollars..it was my previous month’s payout..just a little over 17k .Dollars Last month..3-5 hours job a day…with weekly payouts..it’s realy the Simple job I have ever Do..I Joined This 7 months ago and now making over hourly 87 .Dollars…Learn. More right Here !!bx589a:➽➽➽➽;➽➽ http://DailyOnIineJobsReportsEmployments3620LogicIceGetPay$98Hour…………………………………..=”>. β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…βœ«β˜…β˜…:::::::!!bx589a:….,.

  25. Linda Brodsky

    What is bubble?