Intuitively, why does liquidity premium contribute to bond yield?

Question

According to the Wikipedia, "The upwards-curving component of the interest yield can be explained by the liquidity premium... Liquidity risk premiums are recommended to be used with longer term investments, where those particular investments are illiquid."

I find it hard to understand when looking at graphs online decomposing the all-in yield to interest risk, credit risk, and liquidity risk. Shouldn't liquidity risk be equal to both buyer and seller thus simply widening the bid/ask instead of moving the fair yield upwards? Why should that be factored in the yield (assuming to be the fair yield in the market). Market structure wise, does it have anything to do with the fact that the US corp bond market is still predominantly a dealer's market?

[EDIT] As Alex C pointed out in the comment. This Wikipedia page might be badly written. But I've found similar conclusions mentioned in a few papers. For example, in this paper, the Author says the following

"We find that liquidity is a key determinant in yield spreads, explaining as much as half of the cross-sectional variation in yield spread levels and as much as twice the cross-sectional variation in yield spread changes than is explained by credit rating effects alone"

Also modifying the title to make the question more specific.

Answer 1

For clarity, I'll use two expressions, "liquidity premium" and "illiquidity premium":

• "Liquidity premium" arises when investors value the liquidity profile of an instrument so much that they are willing to pay for the enhanced liquidity, thus pushing the price of the instrument above fair value (and its yield below fair value).
• "Illiquidity premium" arises as ex-ante compensation for expected future difficulty to unwind. Here, "difficulty" can range from higher execution cost, to being forced to sell at depressed prices, to outright inability to sell at all. If you expect these conditions might occur in the future, a rational, risk-averse investor would demand more compensation upfront, in the form of a higher yield (or a lower price).

To focus purely on (il)liquidity premium and not be bothered by credit risk, let's look at examples from the US Treasury market.

First, the chart below shows the yields of all Treasuries in February 2000, chosen for dramatic effects. The most recently issued 10- and 30-year bonds clearly stand out with notably lower yields than surrounding issues. This is partially because these bonds could be financed at lower repos (known as "financing advantage" - this is beyond the scope of this post), but also because investors valued the superior liquidity of these bonds so much that they were literally willing to pay a higher price (and receive a lower yield) for the privilege of buying and holding these bonds.

Let's move to a more stressful environment. The picture below shows the Treasury yield curve on December 15, 2008, the height of the financial crisis. As you can see, investors were once again extremely willing to pay extra for the most liquid bonds (e.g., 10s and CTD into the bond futures), which pushed down their yields significantly. By contrast, some of the older 30-year notes became extremely undesired and very difficult to trade; some investors, in desperate need for liquidity, were forced to sell them at a steep discount (assuming they could find any buyer), pushing up their yields well above fair values.

A more persistent example comes from the Treasury Inflation-Protected Securities market. For many years after the Treasury first started issuing TIPS, the yields of these securities were much higher than justified, mostly because investors demanded an illiquidity premium ("illiquidity" is a strong word in this case; TIPS were liquid, but much less liquid than comparable nominal Treasuries). As it turns out, investors' caution was warranted. During the financial crisis, the TIPS market became extremely illiquid and TIPS traded at very depressed prices, as can be seen in the sharply rising TIPS yield below:

As to whether liquidity premium pushes yield curve to be upward sloping, that's certainly not much of a factor for the Treasury market. A better theory is the existence of term premium, aka bond risk premium.

Answer 2

I do not find convincing the argument that the yield curve is upward sloping due to the lack of a secondary market for longer dated securities. In fact, there is a highly liquid market for 2yr, 5yr, 10yr and 30yr Treasuries and yet the yield curve is still biased to be upward sloping. Intuitively I find that the slope is due to the extra yield that an investor must be paid for investing cash for an extended period of time (yes, an individual investor can liquidate the bonds, but someone must still be holding them). Equivalently, one can think of it as the extra yield being paid for the investor's additional price risk on a long dated security. I think this second effect is what most authors describe as the liquidity effect - the fact that your cash is no longer liquid because it is invested in a long term bond.

Answer 3

This question mixes different concepts. The Wiki page produces all flavours of red herrings. LP is a concept widely used in illiquid credit and alternatives, such as PFI and infrastructure, or other structured credit assets that are infrequently transacted.

1. Term Premium: the additional spread provided (above the expectation for Central Bank deposit rate) to compensate an investor for duration risk.

2. Illiquidity Premium (often called Liquidity Premium): the additional spread provided to compensate an investor in an illiquid credit instrument arising from:

   • asset origination complexity, including specialist skills to source, analyse and structure
   • complexity and regulatory treatment, both of which might affect an asset’s relative appeal
   • a wider bid-offer that reflects challenge to dispose of asset ( ... )
   • not-transparent credit analysis e.g. a private corporate borrower or PFI
   • other instrument features e.g. non-standard coupon

Strictly speaking, any premium from investing in alternative credits may not reflect a reward purely for additional illiquidity risk.

3. Liquidity Risk Premium: the risk and associate cost of closing a position, whether it be liquid or illiquid instrument. This is a risk premium. It is a highly dynamic cost. The cost would be a function of:
   • bid-offer, conditional on the market environment e.g. even UST bills will become illiquid in a crisis (2017).
   • episodic demand for illiquid instrument. This encompasses hard assets such as infrastructure, which are 'known illiquid'. It significantly includes supposedly liquid securities that fundamentally behave illiquidly due to: subdued seasonal turnover, distorted flow and investor appetite, or idiosyncratic investor behaviours.. that may result in a 'dog' i.e. 'unknown illiquidity' factors.
   • illiquid price measurement: circular causation of 'unknown illiquidity' is opaque transaction costs. With a demise of broker dealer liquidity and without regulatory disclosure like TRACE, it is likely that credit securities are traded through a C2C, investor-facing inter-dealer broker. Then, there is no obligation for any parties to report trades, mark instrument prices or otherwise disseminate the price-setting mechanism that occurs at transaction. This opaque behaviour leads to opaque prices.

Answer 4

Consider two bonds each consisting of $N$ annual payments of a coupon $c$ followed by a final payment of $F$ after $N$ years. If an investor (you?) is willing to pay $P_A$ for that promise, then the yield on the bond is $R_A$ defined by
$$P_A = \frac{c}{1+R_A}+\frac{c}{(1+R_A)^2}+...+\frac{c}{(1+R_A)^N}+\frac{F}{(1+R_A)^N}. $$ Now suppose bond "A" has a well-developed secondary market, so that it can be liquidated easily before maturity; and bond "B" having the same coupons and face payment cannot be liquidated at all, and must be carried by the investor to maturity. If you're like most people, the description of bond "B" is less appealing to you, and you would simply be willing to pay a bit more to own bond "A". So if $P_B<P_A$, then you find that $R_B>R_A$ (by the equation analogous to the one above), and that's all there is to it. Empirically, if the bonds appear to be similar in respects other than "liquidity", a price difference (equivalently, yield spread) may be described as a "liquidity premium".