BUFFETT PARTNERSHIP. LTD.
610 KIEWIT PLAZA
OMAHA, NEBRASKA 68131
February 25th, 1970
To My Partners:
This letter will attempt to provide a very elementary education regarding tax-exempt bonds with emphasis on the types and maturities of bonds which we expect to help partners in purchasing next month. If you expect to use our help in the purchase of bonds, it is important that you carefully read (and, if necessary , reread) this letter as it will serve as background for the specific purchases I suggest. If you disagree with me as to conclusions regarding types of bonds or maturities (and you would have been right and I would have been wrong if you had disagreed with me on the latter point either one or two years ago), you may well be correct, but we cannot be of assistance to you in the purchase of bonds outside our area. We will simply have our hands full concentrating in our recommended area, so will be unavailable to assist or advise in the purchase of convertible bonds, corporate bonds or short term issues.
I have tried to boil this letter down as much as possible. Some of it will be a little weighty – some a little over-simplified. I apologize for the shortcomings in advance. I have a feeling I am trying to put all the meat of a 100 page book in 10 pages – and have it read like the funny papers.
I am sure you understand that our aid in the purchase of bonds will involve no future assistance regarding either these specific bonds or general investment decisions. I want to be available at this time to be of help because of the unusual amount of cash you have received in one distribution from us. I have no desire to be in the investment counseling business, directly or indirectly, and will not be available for discussion of financial matters after March 31st.
The mechanics of Tax-Free Bonds.
For those who wish our help, we will arrange the purchase of bonds directly from municipal bond dealers throughout the country and have them confirm sale of the bonds directly to you. The confirmation should be saved as a basic document for tax purposes. You should not send a check to the bond dealer since he will deliver the bonds to your bank, along with a draft which the bank will pay by charging your account with them. In the case of bonds purchased in the secondary market (issues already outstanding), this settlement date will usually be about a week after confirmation date whereas, on new issues, the settlement date may be as much as a month later. The settlement date is shown plainly on the confirmation ticket (in the case of new issues this will be the second and final ticket rather than the preliminary “when issued” ticket), and you should have the funds at your bank ready to pay for the bonds on the settlement date. If you presently own Treasury Bills, they can be sold on a couple of days notice by your bank upon your instructions, so you should experience no problems in having the money available on time. Interest begins to accrue to you on the settlement date, even if the bond dealer is late in getting them delivered to your bank.
Bonds will be delivered in negotiable form (so-called “bearer” form which makes them like currency) with coupons attached. Usually the bonds are in $5,000 denominations and frequently they can be exchanged for registered bonds (sometimes at considerable expense and sometimes free-it depends upon the terms). Bonds in registered form are nonnegotiable without assignment by you, since you are the registered owner on the Transfer Agent’s books. Bonds trade almost exclusively on a bearer basis and it is virtually impossible to sell registered bonds without converting them back into bearer form. Thus, unless you are going to own great physical quantities of bonds. I recommend keeping bonds in bearer form. This means keeping them in a very safe place and clipping the coupons every six months. Such coupons, when clipped, can be deposited in your bank account just like checks. If you have $250,000 in bonds, this probably means about fifty separate pieces of paper ($5,000 denominations) and perhaps six or eight trips a year to the safe deposit section to cut and deposit coupons.
It is also possible to open a custody account with a bank where, for a fairly nominal cost, they will keep the bonds, collect the interest and preserve your records for you. For example, a bank will probably perform the custodial service for you for about $200 a year on a $250,000 portfolio. If you are interested in a custodial account, you should talk to a Trust Officer at your commercial bank as to the nature of their services and cost. Otherwise, you should have a safe deposit box.
The interest received upon the deposit of coupons from tax-free bonds is, of course, free from Federal Income Taxes. This means if you are at a 30% top Federal Income Tax bracket, a 6% return from tax-free bonds is equivalent to about 8-1/2% from taxable bonds. Thus, for most of our partners, excluding minors or some retired people, tax-free bonds will be more attractive than taxable bonds. For people with little or no income from wages or dividends, but with substantial capital, it is possible that a combination of taxable bonds (to bring taxable income up to about the 25% or 30% bracket) plus tax-free bonds will bring the highest total after-tax income. Where appropriate, we will work with you to achieve such a balance.
The situation in respect to State Income Taxes is more complicated. In Nebraska. where the State Income Tax is computed as a percentage of the Federal Income Tax, the effect is that there is no state tax on interest from tax-free bonds. My understanding of both the New York and California law is that tax-free bonds of entities within the home state are not subject to State Income Tax, but tax-free bonds from other states are subject to the local State Income Tax. I also believe that the New York City Income Tax exempts tax-free bonds of entities based within the State of New York, but taxes those from other states. I am no expert on state income taxes and make no attempt to post myself on changes taking place within the various states or cities. Therefore, I defer to your local tax advisor, but simply mention these few general impressions so that you will be alert to the existence of a potential problem. In Nebraska there is no need to have any local considerations enter into the after-tax calculation. Where out-of-state issues are subject to local taxation, the effective cost of your State or Municipal Income Tax is reduced by the benefit received from deducting it on your Federal Income Tax return. This, of course, varies with the individual. Additionally, in some states there are various taxes on intangible property which may apply to all tax-free bonds or just those of out-of-state entities. There are none of these in Nebraska, but I cannot advise on the other states.
When bonds are bought at a discount from par and later are sold or mature (come due and get paid), the difference between the proceeds and cost is subject to capital gain or loss treatment. (There are minor exceptions to this statement as, unfortunately, there are to most general statements on investments and taxes but they will be pointed out to you should they affect any securities we recommend). This reduces the net after-tax yield by a factor involving the general rate of future capital gains taxes and the specific future tax position of the individual. Later on, we will discuss the impact of such capital gains taxes in calculating the relative attractiveness of discount bonds versus “full coupon” bonds.
Finally, one most important point. Although the law is not completely clear, you should probably not contemplate owning tax-free bonds if you have, or expect to have, general purpose bank or other indebtedness.
The law excludes the deductibility of interest on loans incurred or continued to purchase or carry tax-free bonds, and the interpretation of this statute will probably tend to be broadened as the years pass. For example, my impression is that you have no problem if you have a mortgage against real property (unless the debt was incurred in order to acquire municipal bonds) in deducting the mortgage interest on your Federal Tax return, even though you own tax-free bonds at the same time. However, I believe that if you have a general bank loan, even though the proceeds were directly used to purchase stocks, a handball court, etc. and the tax-free bonds are not used for security for the loan, you are asking for trouble if you deduct the interest and, at the same time, are the owner of tax-free bonds. Therefore, I would pay off bank loans before owning tax-free bonds, but I leave detailed examination of this question to you and your tax advisor. I merely mention it to make you aware of the potential problem.
Tax-free bonds are materially different from common stocks or corporate bonds in that there are literally hundreds of thousands of issues, with the great majority having very few holders. This substantially inhibits the development of close, active markets. Whenever the City of New York or Philadelphia wants to raise money it sells perhaps twenty, thirty or forty non-identical securities, since it will offer an issue with that many different maturities. A 6% bond of New York coming due in 1980 is a different animal from a 6% bond of New York coming due in 1981. One cannot be exchanged for the other, and a seller has to find a buyer for the specific item he holds. When you consider that New York may offer bonds several times a year, it is easy to see why just this one city may have somewhere in the neighborhood of 1,000 issues outstanding. Grand Island, Nebraska may have 75 issues outstanding. The average amount of each issue might be $100,000 and the average number of holders may be six or eight per issue. Thus, it is absolutely impossible to have quoted markets at all times for all issues and spreads between bids and offers may be very wide. You can’t set forth in the morning to buy a specific Grand Island issue of your choosing. It may not be offered at any price, anywhere, and if you do find one seller, there is no reason why he has to be realistic compared to other offerings of similar quality. On the other hand, there are single issues such as those of the Ohio Turnpike, Illinois Turnpike, etc. that amount to $200 million or more and have thousands of bondholders owning a single entirely homogeneous and interchangeable issue. Obviously, here you get a high degree of marketability.
My impression is that marketability is generally a function of the following three items, in descending order of importance: (1) the size of the particular issue; (2) the size of the issuer (a $100,000 issue of the State of Ohio will be more marketable than a $100,000 issue of Podunk, Ohio); and (3) the quality of the issuer. By far the most sales effort goes into the selling of new issues of bonds. An average of over $200 million per week of new issues comes up for sale, and the machinery of bond distribution is geared to get them sold, large or small. In my opinion, there is frequently insufficient differential in yield at time of issue for the marketability differences that will exist once the initial sales push is terminated. We have frequently run into markets in bonds where the spread between bid and asked prices may get to 15%. There is no need to buy bonds with the potential for such grotesque markets (although the profit spread to the dealer who originally offers them is frequently wider than on more marketable bonds) and we will not be buying them for you. The bonds we expect to buy will usually tend to have spreads (reflecting the difference between what you would pay net for such bonds on purchase and receive net on sale at the same point in time) of from 2% to 5%. Such a spread would be devastating if you attempted to trade in such bonds, but I don’t believe it should be a deterrent for a long-term investor. The real necessity is to stay away from bonds of very limited marketability – which frequently are the type local bond dealers have the greatest monetary incentive to push.
Specific Areas of Purchase
We will probably concentrate our purchases in the following general areas:
(1) Large revenue-producing public entities such as toll roads, electric power districts, water districts, etc. Many of these issues possess high marketability, are subject to quantitative analysis, and sometimes have favorable sinking fund or other factors which tend not to receive full valuation in the market place.
(2) Industrial Development Authority bonds which arise when a public entity holds title to property leased to a private corporation. For example, Lorain, Ohio holds title to an $80 million project for U.S. Steel Corp. The Development Authority Board issued bonds to pay for the project and has executed a net and absolute lease with U.S. Steel to cover the bond payments. The credit of the city or state is not behind the bonds and they are only as good as the company that is on the lease. Many top-grade corporations stand behind an aggregate of several billion dollars of these obligations, although new ones are being issued only in small amounts ($5 million per project or less) because of changes in the tax laws. For a period of time there was a very substantial prejudice against such issues, causing them to sell at yields considerably higher than those commensurate with their inherent credit standing. This prejudice has tended to diminish, reducing the premium yields available, but I still consider it a most attractive field. Our insurance company owns a majority of its bonds in this category.
(3) Public Housing Authority Issues for those of you who wish the very highest grade of tax-free bonds. In effect, these bonds bear the guarantee of the U.S. Government, so they are all rated AAA. In states where local taxes put a premium on buying in-state issues, and I can’t fill your needs from (1) and (2) , my tendency would be to put you into Housing Authority issues rather than try to select from among credits that I don’t understand. If you direct me to buy obligations of your home state, you should expect substantial quantities of Housing Authority issues. There is no need to diversify among such issues, as they all represent the top credit available.
(4) State obligations of a direct or indirect nature.
You will notice I am not buying issues of large cities. I don’t have the faintest idea how to analyze a New York City, Chicago, Philadelphia, etc. (a friend mentioned the other day when Newark was trying to sell bonds at a very fancy rate that the Mafia was getting very upset because Newark was giving them a bad name). Your analysis of a New York City – and I admit it is hard to imagine them not paying their bills for any extended period of time – would be as good as mine. My approach to bonds is pretty much like my approach to stocks. If I can’t understand something, I tend to forget it. Passing an opportunity which I don’t understand – even if someone else is perceptive enough to analyze it and get paid well for doing it – doesn’t bother me. All I want to be sure of is that I get paid well for the things I do feel capable of handling – and that I am right when I make affirmative decisions.
We will probably tend to purchase somewhere between five and ten issues for most of you. However, if you wish to limit me to your home state, it may be fewer issues – and perhaps those will only be Housing Authorities. We will try not to buy in smaller than $25,000 pieces and will prefer larger amounts where appropriate. Smaller lots of bonds are usually penalized upon resale, sometimes substantially. The bond salesman doesn’t usually explain this to you when you buy the $10,000 of bonds from him, but it gets explained when you later try to sell the $10,000 to him. We may make exceptions where we are buying secondary market issues in smaller pieces – but only if we are getting an especially good price on the buy side because of the small size of the offering.
We will not buy bonds where the issuer of the bonds has a right to call (retire) the bonds on a basis which substantially loads the contract in his favor. It is amazing to me to see people buy bonds which are due in forty years, but where the issuer has the right to call the bonds at a tiny premium in five or ten years. Such a contract essentially means that you have made a forty year deal if it is advantageous to the issuer (and disadvantageous to you) and a five year deal if the initial contract turns out to be advantageous to you (and disadvantageous to the issuer). Such contracts are really outrageous and exist because bond investors can’t think through the implications of such a contract form and bond dealers don’t insist on better terms for their customers. One extremely interesting fact is that bonds with very unattractive call features sell at virtually the same yield as otherwise identical bonds which are non-callable.
It should be pointed out that most Nebraska bonds carry highly unfair call provisions. Despite this severe contractual disadvantage, they do not offer higher yields than bonds with more equitable terms.
One way to avoid this problem is to buy bonds which are totally non-callable. Another way is to buy discount bonds where the right of the issuer to call the bond is at a price so far above your cost as to render the possible call inconsequential. If you buy a bond at 60 which is callable at 103, the effective cost to you of granting the issuer the right to prematurely terminate the contract (which is a right you never have) is insignificant. But to buy a bond of the Los Angeles Department of Water and Power at 100 to come due at 100 in 1999 or to come due at 104 in 1974, depending on which is to the advantage of the issuer and to your disadvantage, is the height of foolishness when comparable yields are available on similar credits without such an unfair contract. Nevertheless, just such a bond was issued in October, 1969 and similar bonds continue to be issued every day. I only write at such length about an obvious point, since it is apparent from the continual sale of such bonds that many investors haven’t the faintest notion how this loads the dice against them and many bond salesmen aren’t about to tell them.
Maturity and the Mathematics of Bonds
Many people, in buying bonds, select maturities based on how long they think they are going to want to hold bonds, how long they are going to live, etc. While this is not a silly approach, it is not necessarily the most logical. The primary determinants in selection of maturity should probably be (1) the shape of the yield curve; (2) your expectations regarding future levels of interest rates and (3) the degree of quotational fluctuation you are willing to endure or hope to possibly profit from. Of course, (2) is the most important but by far the most difficult upon which to comment intelligently.
Let’s tackle the yield curve first. When other aspects of quality are identical, there will be a difference in interest rates paid based upon the length of the bond being offered. For example, a top grade bond being offered now might have a yield of 4.75% if it came due in six or nine months, 5.00% in two years, 5.25% in five years, 5.50% in ten years and 6.25% in twenty years. When long rates are substantially higher than short rates, the curve is said to be strongly positive. In the U. S. Government bond market, rates recently have tended to produce a negative yield curve; that is, a long term Government bond over the last year or so has consistently yielded less than a short term one. Sometimes the yield curve has been very flat, and sometimes it is positive out to a given point, such as ten years, and then flattens out. What you should understand is that it varies, often very substantially, and that on an historical basis the present slope tends to be in the high positive range. This doesn’t mean that long bonds are going to be worth more but it does mean that you are being paid more to extend maturity than in many periods. If yields remained constant for several years, you would do better with longer bonds than shorter bonds, regardless of how long you intended to hold them.
The second factor in determining maturity selection is expectations regarding future rate levels. Anyone who has done much predicting in this field has tended to look very foolish very fast. I did not regard rates as unattractive one year ago, and I was proved very wrong almost immediately. I believe present rates are not unattractive and I may look foolish again. Nevertheless, a decision has to be made and you can make just as great a mistake if you buy short term securities now and rates available on reinvestment in a few years are much lower.
The final factor involves your tolerance for quotational fluctuation. This involves the mathematics of bond investment and may be a little difficult for you to understand. Nevertheless, it is important that you get a general grasp of the principles. Let’s assume for the moment a perfectly flat yield curve and a non-callable bond. Further assume present rates are 5% and that you buy two bonds, one due in two years and one due in twenty years. Now assume one year later that yields on new issues have gone to 3% and that you wish to sell your bonds. Forgetting about market spreads, commissions, etc. , you will receive $1,019.60 for the original two year $1,000 bond (now with one year to run) and $1,288.10 for the nineteen year bond (originally twenty years). At these prices, a purchaser will get exactly 3% on his money after amortizing the premium he has paid and cashing the stream of 5% coupons attached to each bond. It is a matter of indifference to him whether to buy your nineteen year 5% bond at $1,288.10 or a new 3% bond (which we have assumed is the rate current – one year later) at $1,000.00. On the other hand, let’s assume rates went to 7%. Again we will ignore commissions, capital gains taxes on the discount, etc. Now the buyer will only pay $981.00 for the bond with one year remaining until maturity and $791.60 for the bond with nineteen years left. Since he can get 7% on new issues, he is only willing to buy your bond at a discount sufficient so that accrual of this discount will give him the same economic benefits from your 5% coupon that a 7% coupon at $1,000.00 would give him.
The principle is simple. The wider the swings in interest rates and the longer the bond, the more the value of a bond can go up or down on an interim basis before maturity. It should be pointed out in the first example where rates went to 3%, our long term bond would only have appreciated to about $1,070.00 if it had been callable in five years at par, although it would have gone down just as much if 7% rates had occurred. This just illustrates the inherent unfairness of call provisions.
For over two decades, interest rates on tax-free bonds have almost continuously gone higher and buyers of long term bonds have continuously suffered. This does not mean it is bad now to buy long term bonds – it simply means that the illustration in the above paragraph has worked in only one direction for a long period of time and people are much more conscious of the downside risks from higher rates than the upside potential from lower ones.
If it is a 50-50 chance as to the future general level of interest rates and the yield curve is substantially positive, then the odds are better in buying long term non-callable bonds than shorter term ones. This reflects my current conclusion and, therefore, I intend to buy bonds within the ten to twenty-five year range. If you have any preferences within that range, we will try to select bonds reflecting such preferences, but if you are interested in shorter term bonds, we will not be able to help you as we are not searching out bonds in this area.
Before you decide to buy a twenty year bond, go back and read the paragraph showing how prices change based upon changes in interest rates. Of course, if you hold the bond straight through, you are going to get the contracted rate of interest, but if you sell earlier, you are going to be subject to the mathematical forces described in that paragraph, for better or for worse. Bond prices also change because of changes in quality over the years but, in the tax-free area, this has tended to be – and probably will continue to be – a relatively minor factor compared to the impact of changes in the general structure of interest rates.
Discount Versus Full Coupon Bonds
You will have noticed in the above discussion that if you now wanted to buy a 7% return on a nineteen year bond, you had a choice between buying a new nineteen year bond with a 7% coupon rate or buying a bond with a 5% coupon at $791.60, which would pay you $1,000.00 in nineteen years. Either purchase would have yielded exactly 7% compounded semi-annually to you. Mathematically, they are the same. In the case of tax-free bonds the equation is complicated, however, by the fact that the $70.00 coupon is entirely tax-free to you, whereas the bond purchased at a discount gives you tax-free income of $50.00 per year but a capital gain at the end of the nineteenth year of $208.40. Under the present tax law, you would owe anything from a nominal tax, if the gain from realization of the discount was your only taxable income in the nineteenth year, up to a tax of over $70.00 if it came on top of very large amounts of capital gain at that time (the new tax law provides for capital gain rates of 35%, and even slightly higher on an indirect basis in 1972 and thereafter for those realizing very large gains.) In addition to this, you might have some state taxes to pay on the capital gain.
Obviously, under these circumstances you are not going to pay the $791.60 for the 5% coupon and feel you are equally as well off as with the 7% coupon at $1,000.00. Neither is anyone else. Therefore, identical quality securities with identical maturities sell at considerably higher gross yields when they have low coupons and are priced at discounts than if they bear current high coupons.
Interestingly enough, for most taxpayers, such higher gross yields over-compensate for the probable tax to be paid. This is due to several factors. First, no one knows what the tax law will be when the bonds mature and it is both natural and probably correct to assume the tax rate will be stiffer at that time than now. Second, even though a 5% coupon on a $1,000.00 bond purchased at $791.60 due in nineteen years is the equivalent of a 7% coupon on a $1,000.00 bond purchased at par with the same maturity, people prefer to get the higher current return in their pocket. The owner of the 5% coupon bond is only getting around 6.3% current yield on his $791.60 with the balance necessary to get him up to 7% coming from the extra $208.40 he picks up at the end. Finally, the most important factor affecting prices currently on discount bonds (and which will keep affecting them) is that banks have been taken out of the market as buyers of discount tax-free bonds by changes brought about in bank tax treatment through the 1969 Tax Reform Act. Banks have historically been the largest purchasers and owners of tax-free bonds and anything that precludes them from one segment of the market has dramatic effects on the supply-demand situation in that segment. This may tend to give some edge to individuals in the discount tax-free market, particularly those who are not likely to be in a high tax bracket when the bonds mature or are sold.
If I can get a significantly higher effective after-tax yield (allowing for sensible estimates of your particular future tax rate possibilities), I intend to purchase discount bonds for you. I know some partners prefer full coupon bonds, even though their effective yield is less, since they prefer to maximize the current cash yield and if they will so advise me, we will stick to full coupon issues (or very close thereto) in their cases.
I intend to be in the office solidly through March (including every Saturday except March 7th) and will be glad to see any partner or talk with him by phone. To aid in scheduling, please make an appointment with Gladys (or me). The only request I make is that you absorb as much as possible of this letter before we talk. As you can see, it would be an enormous problem if I had to explain each item to all of you.
If you decide you want us to help you in buying bonds, you should let us know:
(1) Whether you want to restrict purchases to your home state for local tax reasons;
(2) Whether you want to restrict us to full coupon issues or let us use our judgment as to where you get the best value;
(3) Your preference as to maturity in the ten to twenty-five year range or if you prefer to let us use our judgment in that area;
(4) How much you want to invest – we may end up several per cent short of the figure you name, but we will never go over;
(5) On what bank the bonds should be drafted.
We will advise you by phone or letter as we buy bonds. Bill and John will be doing much of the mechanical work. Needless to say, none of us will have any financial interest in any transaction. Should you have any questions regarding the mechanics, please direct them to John or Bill as I will probably be swamped and they will be more familiar with specific transactions. After March 31st, I don’t expect to be around the office for several months. Therefore, if you want to talk things over, come in by then. The completion of all purchases may go into April, but Bill will be taking care of this and the mechanics will all be set up.
You should realize that because of the enormous diversity of issues mentioned earlier, it is impossible to say just what will be bought. Sometimes the tax-free bond market has more similarities to real estate than to stocks. There are hundreds of thousands of items of varying comparability, some with no sellers, some with reluctant sellers and some with eager sellers. Which may be the best buy depends on the quality of what is being offered, how well it fits your needs and the eagerness of the seller. The standard of comparison is always new issues where an average of several hundred million dollars worth have to be sold each week – however, specific secondary market opportunities (issues already outstanding) may be more attractive than new issues and we can only find out how attractive they are when we are ready to make bids.
Although markets can change, it looks as if we will have no difficulty in getting in the area of 6-1/2% after tax (except from Housing Authority issues) on bonds in the twenty-year maturity range.
Warren E. Buffett