Sunday, February 16, 2014

The cost of manufacturing liquidity premia

I'm going to use a stock market analogy to work out the costs of manufacturing liquidity premia.

In addition to paying dividends and providing price appreciation, stocks provide an amenity flow in the form of expected exchangeability, or liquidity. Like any other consumption good & service, the provision of such an amenity requires an outlay by the supplier. In the same way that a widget producer won't sell widgets below marginal cost, the marginal seller of stock, the issuing firm, won't manufacture new liquidity if the cost of production exceeds the benefits.

Say that a firm can hire an entire investor relations department for next to no cost. The IR team, full of eager promoters, will double the price that the marginal investor is willing to pay for the liquidity services thrown off by the firm's stock. They go about improving the stock's liquidity services by evangelizing the firm's "story", thereby widening the base of investors who deal in the shares. They make it easier for investors to hop in and out of the market.

The firm can now issue new stock at a much higher price thanks to its swollen liquidity premium. It invests the proceeds of new issuance at the risk-free rate of return. The firm has succeeded in getting something for nothing—at no cost to itself it has increased earnings-per-share. By issuing more shares the firm can continue to earn these extra-normal returns, at least until new issuance has satiated investor demand for liquidity and driven the firm's liquidity premium back to its previous level, at which point the strategy will have exhausted itself. New stock issuance will no longer increase per-share earnings.

Of course, investor relations teams can't be hired for nothing. If firms could perpetually increase earnings per share by costlessly boosting their liquidity premium and issuing new shares, then everyone would be doing it. In general, the price of hiring an IR team should be set such that it just offsets the benefits of the increase in a firm's liquidity premium. If the cost is lower, then firms will all try to purchase IR services in order to enjoy extra-normal profits, driving IR costs higher until the window for extra-normal profits has been closed. If the cost is higher, then firms will fire their IR department, the benefits of the liquidity premium manufactured by the IR department not justifying the expense of paying their salaries. In this case, IR costs will retreat until firms once again see some advantage in trying to hire an IR department to generate liquidity premia.

In short, the price that investors pay to enjoy a liquidity premium will be competed down to the IR costs of producing that premium.

Incurring IR costs are not the only way to do encourage liquidity. Relationships with market makers, dealers, and investment banking research departments will also help boost liquidity premia. Illegal practices like wash-trading can do the trick too.

Keep in mind that there are also large network effects at play. Incumbent stocks that have enjoyed high liquidity premia for decades will remain locked into that position, even if the incumbent's CEO were to fire the entire IR department. Liquidity is sticky. It would take incredibly large marketing outlays for a small rookie stock to displace an incumbent like IBM from its superior liquidity position.

If you get the chance, try visiting ten or twenty websites of publicly-traded companies and note how fancy each IR section is. Some will have only bare-bones text (like Berkshire Hathaway), others will have incredibly fancy flash animation and gorgeous pictures (like any junior gold miner).

These websites will give you some sense for each firm's pool of potential projects and respective strategy. Firms with a plenty of high-yielding investment opportunities will see no benefit in allocating funds to boost their liquidity premium. These sorts of firms will generally have ugly IR sections. Firms with fancy IR websites, on the other hand, have presumably measured all potential investment opportunities and decided that the highest yielding one is to hire aggressive IR so as to boost their liquidity premium, subsequently floating new shares. Investors who value liquidity on the margin would do well to gravitate to firms that see value in manufacturing liquidity premia. If you want liquidity, then buy from the people who make the stuff. Those investors who prefer pecuniary returns rather than liquidity returns should always avoid firms with fancy IR pages. After all, why buy liquidity if you don't value it?

The boundary case of firms pursuing higher liquidity premia are penny stock promotes. These firms have no real underlying business. Their only purpose is to create a temporary liquidity premia, the insiders exiting before those premia collapse to zero.

The general principles behind the manufacturing of liquidity premia described above apply not just to stocks, but to bonds, bitcoin, gold, cars, land, banking deposits, and all sorts of other assets. As long as people face uncertainty, they will always want to own liquidity. Those skilled in the manufacturing of  this liquidity—marketers, salespeople, investor relations execs, promoters, bankers, and evangelizers—will always find their talents in high demand.


  1. I thought you were talking about banks. They manufacture liquidity.

    1. I specifically used stocks as my focus point. You don't think it works?

    2. JP: It works. But I did read it thinking it was one long metaphor, expecting you to suddenly say: "Aha! I'm talking about banks!"

    3. Ah, I see. You're right that this post is a metaphor of sorts (more of a template perhaps?). The basic principles that seem to apply to equity markets can be easily transferred over to other asset markets, including banks. (I've updated my conclusion to mention banking deposits/bankers).

  2. Nick, JP,
    In that case, is liquidity specifically different from volume and velocity?

  3. Objective characteristics of an asset's secondary market such as volume, velocity, bid ask spreads, market depth, etc. engender a certain degree of liquidity in an asset. The price that people will pay to enjoy that collection of features is the liquidity premium.

  4. Those investors who prefer pecuniary returns rather than liquidity returns should always avoid firms with fancy IR pages.

    Hmm, interesting idea. Assuming this relationship actually holds, I wonder how long it will it will take for long-term investors to bid up the prices of ugly IR page companies?

    A possibly dumb question: you may have addressed this before, but why is the stock of Google worth so much if, as its FAQ maintains, Google has no intention of ever paying a dividend? In the event that Google is ever wound up, its assets would presumably be worth a lot less than $300-400 billion. I guess "expected future price appreciation" is an answer, but what is the source of price appreciation w/o any future expected dividend payments?

    (To be precise, Google's FAQ says no plans for a dividend in the unspecified "foreseeable future." But assume Google were absolutely, credibly committed to paying no dividends, ever. Should its stock rationally equal only the estimated PV of its assets in the event of bankruptcy?)

    1. John S. Remember that profits can be used to buy back shares, and buying back shares is equivalent to paying dividends (except for taxes).

    2. I agree with Nick. If income isn't paid out in the form of dividends, it accumulates at the firm and will be reinvested. As a shareholder, you still have a claim on those resources. You'll get access to them someday in the form of a buy back, take-over, or a windup.

    3. Right, I overlooked buybacks. A couple of questions:

      1. Why don't more companies offer variable dividends instead of fixed nominal amounts? (Does any firm do this?)** Buybacks are another way to return money to shareholders (with tax benefits), but they're rather unpredictable events. Wouldn't some firms/investors like the chance to hedge against losses/gain from profit windfalls with a bit more predictability?

      **The only examples I could find are Progressive Corp. and Cal-Maine Foods.

      2. Stock buybacks reminds me of the whole central bank = blowfish, fiat money = bubble, gold --> CPI units discussion a while back. After skimming some old posts, I get the feeling that Nick's answer only reveals another Russian doll. As summarized by JP:

      "The redemption rate is no longer fixed. Rather, it is adjusted to ensure that, in the secondary market for central bank money, that money's value relative to goods-in-general falls by 1-3% a year."

      But w/o direct redemption of baskets, what is the actual thing/mechanism that connects central bank money's value to the CPI basket? Somalian/Iraqi fiat memory, Selgin's rocket/satellite metaphor, network effects, tax-based chartalism? To say the central bank maintains the value of money by maintaining the value against a CPI basket seems like a merry-go-round.

      3. JP has said that the central banks assets "back" the value of money, HT Mike Sproul. But since (at least in the US until 2009), the central bank's assets primarily consist of government bonds which promise to pay fiat money, aren't we back at the beginning of the "bubble" merry-go-round?

    4. 1. In theory, dividend policy is irrelevant.

      2. The actual mechanism that connects bank money's value to the CPI is open market operations, which by necessity requires that something be on the issuer's balance sheet. Or the payment of interest, which requires assets to generate the necessary funds to pay interest. The return of assets upon wind up is distant mechanism.

      3. Yes, it's a weird sort of equilibrium. Highly unstable, in theory at least. If assets back money, I'd expect to see more volatility in prices over time than we actually tend to see.

    5. On the stability of dividend policy, it mostly has to do with governance issues I believe. In countries where financial regulation is weaker and corporate boards have little power, dividend policy is usually even less variable. It's a costly signal from management to prove they aren't screwing shareholders over at the benefit of themselves or insider trader friends.

    6. 1. JP, thanks for the link. Some quick googling reveals the same points made by John Hawkins about signalling (thank you, as well). Personally, my intution re: dividends leans towards Lintner's "bird in the hand theory," though I wasn't familiar with the name.

      This was an enjoyable summary of the issue for me. "Is the Dividend Puzzle Solved?"

      2, 3. I realize you're busy now, so I won't pursue these lines of questioning further for now; I'll just read your relevant posts (and Nick's) more carefully. I did like this part of your comment on Nick's CPI standard post:

      "It seems to me that you could substitute the "central bank money" in your story with any other private liability (or even equity) and the story is still true. A private company could issue bearer notes initially 100% backed and convertible into gold and then go through all the permutations you've listed to arrive at a note linked to CPI"

      I had a thought about private chartal currencies like XRP. Wouldn't a large corporation like Google be much better situated than Ripple Labs to launch a real-world, stable value chartal currency since Google can issue new shares to mop up excess "notes" (private OMOs)? If the value of G-bucks falls too low (relative to a CPI basket), Google could increase the demand for G-bucks (and reduce G-buck supply) by selling new issues at a favorable price to G-buck holders.

      (I'm probably missing some arbitrage trick here that would crash the idea).

    7. FWIW, Google owns a chunk of Ripple Labs.

      However, when it comes to Ripple I still think it's the IOU side that's interesting, not XRP.

    8. True. Maybe someday Google will integrate Wallet with Ripple to avoid credit/debit card fees.

      I think both IOUs and XRP are interesting. Both are doing something different in the CC space besides BTC-style mining, which doesn't seem sustainable in the long-run.

  5. I mostly agree with what you're saying, but a few things strike me.

    1. Much of what IR does is to try to change expected pecuniary returns, as opposed to changing expected liquidity returns. In practice, it may be hard to separate the two effects.

    2. Improving expected liquidity returns may increase the share price, but it's never going to outweigh the risk premium investors need, to the extent that the company can afford to reinvest at the risk-free rate, is it? (At least not on an indefinite basis).

    3. Even with the no-cost IR, I'm not sure the break-even is when the liquidity premium has gone back to its original level, simply because there are other factors at play. As stock levels rise, investors my start to expect fewer profitable investment opportunities (especially if the company is investing in risk-free assets). So falling expected pecuniary returns may cancel the additional liquidity benefit.

    1. Nick, your point on IR doing more than promoting liquidity is well-taken.

      On #2 and 3, here's a bit of colour. When I originally wrote this I was imagining a small firm that invested all funds in government bonds. Investors in the firm wouldn't require a risk premium since buying shares would be the same as buying government bonds. An investment in no-cost IR would drive the share price higher, allowing the firm to increase its per-share exposure to government bonds via new issuance, thus boosting EPS.

  6. John S.
    "3. JP has said that the central banks assets "back" the value of money, HT Mike Sproul. But since (at least in the US until 2009), the central bank's assets primarily consist of government bonds which promise to pay fiat money, aren't we back at the beginning of the "bubble" merry-go-round?"

    As long as the central bank holds some real assets as an anchor, it is possible to back the dollar with bonds that are themselves claims to dollars. A stock market analogy: GM issues new shares of stock and uses them to buy call options on GM stock (issued by Merrill Lynch). This will make GM stock more volatile. A rise in GM's share price will make the calls more valuable, but since those calls are GM's asset, that will make the stock more valuable, which will raise the calls still more, etc.

    Monetary example: A bank receives 100 oz of silver on deposit, against which it issues 100 paper receipts ('dollars'). At this point, $1=1 oz. Then the bank issues another $200 and uses them to buy $200 worth of bonds (denominated in dollars). Define E as the value of the dollar (oz/$). Setting assets (100 oz + $200 of bonds) equal to liabilities ($300 of bank notes) yields:


    or E=1 oz/$. Thus the bank can issue new dollars, backed by dollar-denominated assets, without affecting the value of the dollar.

    But now suppose the bank is robbed of 30 oz. (10% of its assets). The equation becomes

    70+200E=100E, or E=0.7oz/$

    So the effect of feedback is that a 10% loss of assets causes a 30% inflation. It's therefore a mistake for central banks to hold assets denominated in their own currency.

    1. Mike, thanks for the example, esp. the call option feedback analogy. I like the backing theory (though it makes my head spin). I'll have to read your posts more carefully.

      Quick question: if it's "a mistake for central banks to hold assets denominated in their own currency," what assets should they hold?

    2. John S:

      Central banks should hold assets NOT denominated in their own currency in order to avoid the inflationary feedback effect. For example, it would be a good idea for the British central bank to hold US bonds instead of British bonds.

      On the other hand, if the Bank of England holds US bonds, then it would be bad for the Fed to hold British bonds. If US bonds lost value, then the pound would lose backing and fall in value. But then British bonds (denominated in pounds) would lose value. Those British bonds were backing the dollar, so the dollar would lose value, and so on.

      It's the quantity theory that makes my head spin. The backing theory is easy. It just says that the dollar has value because of the assets backing it, just like every other financial security. The quantity theory has to assert that paper money is valued on principles that are different from the principles governing every other financial security.

    3. Ok, you've given me examples of assets that central banks shouldn't hold. If you were Fed dictator for life, what composition of assets would you strive for, roughly? (There don't seem to be enough foreign bonds, which aren't themselves partially tied to the value of the dollar, to back the currency of the US).

      I don't mean to badger you, I'm just genuinely curious.

    4. John S:

      I'd have the Fed hold a diversified portfolio of land, bonds, metals, commodities, and short term real bills, with the specification that the Fed holds physical amounts of those things (e.g., ounces of gold, acres of land) rather than a given dollar's worth of those things. In the case of bonds, they should be inflation-protected.

      Feel free to badger any time.

    5. What about shares in broad-based index funds like the Vanguard Total Stock Market Index Fund, or its foreign equivalents? Perhaps the specification should be "% of mkt cap" of all global stock markets.

    6. Depends on what the public wants. If they don't mind that the value of their bank notes might fluctuate with the stock market, then a stock fund could be included. But since money-issuing banks have historically taken pains to keep their money stable, stock funds seem unlikely to be a preferred investment.

      I should also mention that when banks have issued money against land, they have commonly insisted that $2 worth of land be posted as collateral for every $1 worth of money issued (on loan). So even if land fluctuates in price, the money backed by that land could remain stable.

    7. Doesn't this depend on one's view on how independent the central bank really is?

      If you take the view that the central bank's backing includes the government's ability to levy taxes, then any actual assets on the central bank's balance sheet is just icing on the cake. It's not going to make much difference to the value of the bank notes, is it? If the central bank is truly on its own, then it might be a different story.

    8. Nick:

      Correct. Money is valued more like bonds than like stocks. If the issuer has $1 mil of assets, and it has issued only $100 of money, then (just like with bonds) the issuer can lose all but $100 of its assets without affecting the value of its money. Stocks, on the other hand, will rise or fall with every change in the issuer's assets.

      So if the government stands ready to bail out the central bank, then the central bank's assets can go all over the place without affecting money's value. The same is true if either the central bank or the government has assets far in excess of the amount of money issued.

      On the other hand, if the central bank (and/or the government) has barely enough assets to cover the money issued, then the value of money will fluctuate in step with the assets backing that money.

    9. Mike Sproul: thanks for the explanation above about using bonds to back a currency. I can understand that one, but not the stock one. I always wondered about that. I find myself telling people of your backing theory idea a lot, even though I've never really taken the time to properly study it (read your papers carefully). Recently Cullen Roche asked me about it on pragcap, and I did my best to fake my way through, but you might want to undo any damage I caused before it's too late: :D

    10. Tom:

      Thanks! I'll check out pragcap in a bit.

      I might have caused some confusion by saying in one place that money can be backed BY stocks, while saying in another that money is valued less LIKE stock than like bonds.

      Money can be backed BY anything of value: by land, metal, stocks, bonds, lottery tickets, etc.

      Stocks and bonds can also be backed BY anything of value, even by other stocks and bonds.

      But when we look at how the value of money is related to its backing, money is more LIKE bonds than like stocks. Money, like bonds, is a (mostly) fixed claim to the issuer's assets, while stocks are an equity claim to the issuer's assets.

      So suppose that the dollar is pegged to one ounce of silver. If a bank has issued $100, then as long as that bank's assets are worth at least 100 oz, then $1=1 oz, and this will still be true if the bank's assets rise to 1 million oz . But once the bank's assets fall below 100 oz, the value of the dollar will fall in step, so that if the bank's assets are worth just 70 oz, then $1=0.7 oz.

  7. JP - Have you ever looked at the differences in yields between on-the-run vs. off-the-run government maturities (on-the-runs are the most recently issued of their maturity and are typically more actively traded, off-the-run are very well season)? It's a very interesting way to look at the differing liquidity premia for different treasury issues.

    1. I'm definitely aware of it. If I had some actual data to make a chart I'd write a post on the topic, since as you point out the difference is comprised of a liquidity premium---my favorite topic.