To analyze trading in newly issued municipal bonds, the Journal obtained the following data sets:This data set includes all municipal-bond trades from April 2013 through December 2017. There were about 28.5 million trades during the period of the analysis.
Each trade record includes: • The bond’s Cusip (Committee on Uniform Security Identification Procedures), a nine-character alphanumeric code used to identify securities including stocks and bonds. • A description of the bond that typically includes the name of the issuer • The interest rate paid on the bond, known as the coupon. • The maturity date, which is the date on which the issuer will repay the face value of the bond to holders. • The type of trade, denoted by one of three letters. (S = Sale to a customer by a dealer. D = Trade between two dealers. P = Purchase from a customer by a dealer.) • The face value, also known as par value, of the traded bond. • The price of the trade. • The yield to maturity, or annual interest earnings, that the bond buyer would receive, expressed as a percentage and calculated based on the price of the trade and the interest rate and maturity date of the bond. The characteristics of bonds issued from 2013 through 2017, from Mergent, Inc. The data documents the characteristics of each bond, including: • Issue ID, created by Mergent, is used to identify separate bonds that were part of the same of bond issue. • Cusip • Offering date • Name of issuer • Lead underwriter(s) and supporting underwriters • Maturity date • Call dates, when the issuer has the option to pay off and retire a bond before the maturity date. • Total face value, or par amount • Interest rate, or coupon • Offering price • Offering yield to maturity, generally the annual interest an investor would earn if the bond is purchased at the offering price. • Whether the bond issue was “competitive,” in which the issuer put the bonds up for auction and awarded them to the underwriter offering the lowest total interest cost, or “negotiated,” in which the issuer selected an underwriter to structure and price the bonds without competition. Bond-trading records of insurance companies, from the National Association of Insurance Commissioners Regulatory reports filed each year by insurers, include disclosures of every trade in which the insurer bought and/or sold bonds during the year. The Journal obtained these records for 2013 through 2017. Each record includes: • Cusip • Date purchased or sold • Face value • Total consideration paid or received • Name of the dealer with which the insurer traded The Municipal Market Data daily index of AAA bond yields, from Refinitiv The MMD index is the primary indicator used to track the municipal-bond market. It estimates the average yield to maturity of recent trades in AAA bonds, with one figure each market day for each maturity from 1 to 30 years. EMMA, a website of the Municipal Securities Rulemaking Board The Journal used the website to access the offering documents of bond issues and to confirm details about individual bonds, including offering price, offering yield and maturity date. Overview of the bond issues
Earliest Issue Date
Latest Issue Date
The analysis The Journal’s preliminary analysis of trading in newly issued bonds found tens of thousands of bonds whose prices rose above their offering levels soon after they started trading. Once those bonds were identified, the Journal determined that the price run-ups often began with early “flipping” of the bonds. Flipping occurs when a trader buys the bond, then quickly sells it for a higher price. The Journal focused on bonds that had at least one trade in which $100,000 or more of par was bought by a customer, then sold to a dealer within the first two days of trading. There were 29,689 bonds that met those criteria. This table summarizes those trades.
Result of sale to dealer
Par purchased from customers first two trading days
Percentage of total offering par purchased from customers first two trading days
Net to sellers over offer price
Net as percentage of purchased par
As the table shows, flips in the first two days of a bond issue are much more common in negotiated offerings than competitive ones, with about 5% of the total par flipped in negotiated offerings, compared with less than a third of a percentage point in competitive ones. It also shows that most of the flips resulted in a trading profit for the customers who bought the bonds and quickly sold them to dealers. In negotiated offerings, those customers made about $291 million on those sales. The Journal then analyzed what happened to the prices of these bonds following the flips. To calculate the prices that buy-and-hold investors ultimately paid for the bonds, the Journal took all the sale-to-customer trades up to the 10th trading day and worked backward from the latest trade until the par amount traded equaled the total par amount of the bond. (Many bonds were fully sold to investors before the 10th trading day.) This method excludes from the price calculation the original sales to customers that were flipped back to dealers. The Journal then calculated the weighted-average price and yield of those trades. The estimated total profit or loss to the initial customers who flipped the bonds and dealers who then sold them to other dealers and other customers would be equal to that calculated price minus the offering price, multiplied by the total par offering amount of the bond. Market adjustment The value of an individual bond fluctuates with changes in the broader market for municipal bonds, generally rising as average market yields decline and rising as they fall. To account for those market forces, the Journal compared changes in the yield to maturity of sales to customers with changes, on the same days those trades occurred, in the MMD curve, which tracks the broader market. This table summarizes the net and market-adjusted trading profits during the first 10 days of trading in the flipped bonds.
Total offering par of bonds that were flipped
Par amount that was flipped during first two trading days
Net markup on sales to investors first 10 trading days
Market-adjusted net markup
Percentage of net markup not attributable to market
Insurer flipping of newly issued bonds The names of the dealers and customers who participated in bond trades aren’t included in the MSRB data, so the Journal couldn’t determine how much of the bond flipping was facilitated by the bond underwriters. The Journal got that information for a small subset of the flips using the NAIC trading records of insurers. Unlike the MSRB data, the insurers’ trading records disclose the names of the dealers with which the insurers traded. The table below shows some of the data points contained in a typical NAIC record from the NAIC data. The table shows that Safety National Casualty Corp. bought bonds issued by the West Contra Costa Unified School District with a total face value of $1 million from Piper Jaffray & Co. (the firm’s name was misspelled) for $969,480 on March 2, 2016. That was the first day the school district’s bonds traded. It also shows that the insurer sold the bonds back to Piper Jaffray that same day for $981,000, a profit of $11,520. Piper Jaffray was the lead underwriter on the district’s bond offering.
Safety National Casualty Corporation
Insurance Co. Parent
Tokio Marine Holdings Inc GRP
W CONTRA COSTA USD-A
NAME OF VENDOR
PIPER JAFFREY & CO
NAME OF PURCHASER
PIPER JAFFREY & CO
PAR VAL BNDS STKS
REAL GN LS ON DISP
Of the roughly $60 billion of bonds in the Journal’s analysis that were flipped within the first two trading days, insurers bought and sold $3.7 billion of the bonds, or about 6%, the NAIC data shows. And 70% of the time, the lead underwriter sold the insurers the bonds and then bought them back. Mattia Landoni, an assistant professor at Southern Methodist University’s Cox School of Business, reviewed the Journal’s methodology and findings.
Copyright ©2019 Dow Jones & Company, Inc. All Rights Reserved. 87990cbe856818d5eddac44c7b1cdeb8