Buffett v the Landesbank – the US gets paid to have a trade deficit

It remains conventional wisdom that the US current account deficit is the accumulation of a debt that will one day need to be repaid. There are very good reasons to believe this is false. The US in fact has net external assets, and this is not a temporary phenomenon – it’s structural.

How can the US constantly run current account deficits and accumulate net external assets? Simply put, by being better at trading assets and investing than the rest of the world. Imagine Warren Buffett buying assets on behalf of America and Germany’s Landesbanken acting on the rest of the world’s behalf. You may feel like stopping there. One invests for the long-term in high-returning, highly efficient businesses, the other falls for every investment fad, has no deep understanding of risk, and systematically destroys value. Assume that each year the Landesbanken invest slightly more than Warren – which one do you think will create more wealth?

That in fact seems to be the story of the US net external balance sheet. Now, some economists get very confused by this observation – mainly because they just take the BEA’s estimates of the US net external balance sheet position at face value. On the BEA’s measures, the US has net external liabilities of around $7trn.

How could they be so wrong? Mainly because their method of measurement is deeply flawed. Typically, external asset/liability positions are proxied by cumulative flows on the current account – so if you run a deficit you have ‘liabilities’ – or through attempts to measure ‘market values’, for example, by proxying changes in oversees FDI valuations using price changes in local stock markets.

These measures are completely misleading, for a host of reasons. For starters, investing in US overseas FDI is likely a far higher returning (and less risky) investment strategy than buying the MSCI world equity market ex the US. Similarly, overseas FDI into the US is notoriously poor – you would almost certainly be better off in the S&P500.

Buffett v the Landesbank

It seems far more sensible to take a present value of net income to measure the net external balance of the United States. Original work on this from the early 2000s has been supported by subsequent trends in US international net income (see chart below), later popularised by the unhelpfully-named ‘dark matter’.

But the key point is that using a present value of net income is only significantly misleading if the riskiness of US external assets is far greater than the risk of foreign holdings in America – or the income received is far more volatile than the income being paid. Risk is a tricky subject, at best. Is a 10-year treasury bond yielding sub-2% more or less risky than an Intel chip plant in Malaysia, which makes a double-digit return on capital? Not obviously. If we simply proxy risk by the skew on the return profile, the treasury is damn risky.

US net international investment income

Another simple way to consider the validity of taking the present value of net income is to look at the relative volatility of the income streams – intriguingly, the income the US receives on its oversees assets appears less volatile than what it pays. In other words, US overseas assets appear less risky than foreign assets held in the US. In which case a simple present value of the US net investment income understates its position – we should apply a higher discount rate to foreign-owned assets in the US than the income from US assets overseas.

So here’s the crux: American generates a high positive net income from its net international asset position – which implies its net external asset position is a positive number. Despite running permanent current account deficits, the US has accumulated net external assets. This sounds less counterintuitive if you think that all it means is that US assets oversees are more valuable than foreign holdings of US assets.

To highlight how significant the difference is, BEA measures of US net external liabilities are around $7trn, if we simply value the net income on a PE multiple of 14x, the net external assets of the US are around $2.7trn!

An inconsistency this large seems extreme, but consider the valuation of foreign direct investment (FDI). According to the BEA, net FDI of the US has a value of less than $1trn, although the US earns $424bn on its overseas FDI and foreign FDI in the US only earns $153bn. If we value both income streams on 14x earnings, the US has net FDI assets of $4.2trn – the BEA ‘market value’ estimates are out by $3trn.

Recent academic research has placed greater emphasis on the point that the US has positive net investment income, while running persistent deficits, but rarely concludes that this is a sustainable trend, or that it amounts to a steady accumulation of net assets. This excellent paper by Gourinchas & Rey instead draws an analogy between the US and a leveraged venture capital firm – the US is ‘borrowing’ by issuing treasury bonds and other ‘safe’ assets and is investing in high risk/high return assets overseas.

This is consistent with a leveraged “cowboy” stereotype. The correct analogy is Warren Buffett v the Landesbank. What looks most plausible is that the US has an edge in trading assets and making superior foreign direct investments. Selling balance of payments “insurance assets” to Asia and EM (i.e. low yielding treasuries) and accumulating high return FDI-assets in the same geographies creates net assets. Don’t forget, insurance has negative expected returns to the buyer.

US portfolio investments are similarly superior – you know when a stock market is cheap or a positive structural break in profits is about to occur when US private equity and long-term investors start accumulating large positions. By contrast, foreign investors bought most US equity at the peak of the tech bubble in the late 1990s – US ‘liabilities’ which subsequently evaporated. And if you think its a one-off, remember how much subprime ended up on the balance sheets of Landesbanken, and other European lenders. Not only did the US sell low yielding pseudo-investment grade insurance. It then wiped out hundreds of billions of liabilities. And guess who is hoovering up non-performing loans in Italy and peripheral Europe at double-digit hurdle rates, as the European regulator forces local banks to engage in fire sales? Which country do the private equity firms with the largest stakes in Irish banks at the share price lows come from?

Now it is true that extreme differential risk preferences could result in wildly different market values for assets and liabilities, even if the net income stream is consistently positive. If a firm has borrowed at extremely low interest rates but has a high stock of debt we would be concerned, because at some point the debt might be redeemed or the refinancing rate might rise. Clearly the nature of US external assets are very different to its liabilities.

At the same time, of course, the balance of payments is not a firm. Having high foreign holdings of treasuries does not pose anything like the financial risk it has obtained amongst the paranoid commentariat. We should all have learnt from Japan and the Eurocrisis that in money printing economies, government bond prices are determined primarily by interest rate expectations. One implication is that it really doesn’t matter what “foreigners” do. At the same time, of course, if US interest rates rise substantially, the profitability of Google’s overseas assets does not rise in tandem, so net income could deteriorate. But I still wouldn’t bet on US overseas investment income not compounding away at a far high rate – nor would I bet on far higher US interest rates any time soon.

Trading assets and selling financial “insurance” is a core US competitive advantage. In today’s world that appears cumulatively more valuable than producing goods. So the economic significance of the US current account deficit to the balance sheet of America is trivial.

The official data completely mis-measure the US net external position – implying that it has net external liabilities. But the true picture is revealed by the net investment income which is significantly positive. Discount that cashflow to create a net present value and the US has significant net assets. The current account deficit is better than a free lunch – America gets paid to have lunch.

Thanks to @JWMason1 @NeilLancastle @Ramanan_V and @JoMicheII for discussion on Twitter

About The Author

Eric Lonergan is a macro hedge fund manager, economist, and writer. His most recent book is Supercharge Me, co-authored with Corinne Sawers. He is also author of the international bestseller, Angrynomics, co-written with Mark Blyth, and published by Agenda. It was listed on the Financial Times must reads for Summer 2020. Prior to Angrynomics, he has written Money (2nd ed) published by Routledge. He has written for Foreign AffairsThe Financial Times, and The Economist. He also advises governments and policymakers. He first advocated expanding the tools of central banks to including cash transfers to households in the Financial Times in 2002. In December 2008, he advocated the policy as the most efficient way out of recession post-financial crisis, contributing to a growing debate over the need for ‘helicopter money’.

28 Responses

  1. Ramanan

    “An inconsistency this large seems extreme, but consider the valuation of foreign direct investment (FDI). According to the BEA, net FDI of the US has a value of less than $1trn, although the US earns $424bn on its overseas FDI and foreign FDI in the US only earns $153bn. If we value both income streams on 14x earnings, the US has net FDI assets of $4.2trn – the BEA ‘market value’ estimates are out by $3trn.”


    Netting is not the right way to do it. When looked at the gross values, it doesn’t seem that wrong.

    US FDI abroad is around $7tn and earns $775bn

    Foreigners’ FDI is $6tn and earns $574 bn.

    There’s nothing abnormal about these numbers. Surely one can make the case that foreigners are just catching up and don’t mind a low earning as compared to outward FDI.

    Netting and calculating returns is wrong mathematics. Suppose the US held $100 of FDI abroad and foreigners held $100 of FDI in the US, and if on net the US earns $1, then you cannot say that the US is earning $1 on nothing or that the return is infinite/b> (because the net stock of FDI is 0).

    Second, your graph should ideally be plotted relative to GDP. As in net interest income/GDP. By plotting it your way, you are misleading.

    More importantly, you have to look at the dynamics. The US CADs keep adding to the liabilities and the interest income will turn negative.

    On the production side of the economy, the term “free lunch” is misleading to the core. Current account imbalance drains demand and output. Only if the US were to have full employment and that being sustainable can you claim that there is some “free lunch”. Plus an attempt to raise demand by whatever means implies a deterioration of the current account, and the ‘tipping point’ will reach even faster. The tipping point has been kept far by a lower domestic demand and output in the US.

    • Eric Lonergan

      Thanks for the comment Ramanan. I come to my conclusion by looking at the matter in gross terms. My question is very simple: why would you value $100 of income paid higher than $100 of income received, particularly if the income received is less volatile. Valuing these two income streams on the same multiple results in a far higher stock of net assets. Bear in mind, the BEA proxies are terrible: US FDI generates returns far higher than overseas stock markets.

      It’s fine to look at the absolute level of income. The point is that the superior investment activity of the US appears structural – because the income keeps rising. If liabilities were being accumulated, the opposite would be happening. Think Buffett and Landesbanken.

      • Eric Lonergan

        Ramanan – I’ve read your post. You do realise that the ‘market value’ approach is not actually using market values? There are no observable market values for US FDI. The BEA is *proxying* the value of FDI using the price returns of local stock markets. This is not done in any corporate accounting – do u think it is likely to produce meaningful results? That’s really the question. I am saying that this proxy is producing nonsense results, as highlighted by the growth and risk properties of net income. This is really the substantive issue – I would be v interested if u have any views on it.

      • Ramanan


        Basics first. There is nothing wrong inherently about the FDI numbers as my example shows. Netting the stock and FDI and income can give bizarre numbers if you take ratios even though the numbers may be fine.

        My post has nothing to do with how they are estimated. I am just arguing that correct numbers may look bad if not analysed correctly such as $1 income on net FDI of $0.

        So netting needs to be done carefully. So your argument doesn’t work.

        About your claims on measurement: have any link?

      • Eric Lonergan

        The problem is Ramanan that the numbers being used for gross and net assets are nonsense. The income data however are much more accurate and are not proxied. The ‘market value’ numbers are usually quoted in external liabilities – see this for methodology http://www.federalreserve.gov/econresdata/notes/feds-notes/2014/measuring-direct-investment-in-the-financial-accounts-of-the-united-states-20141031.html

        As you can see different methodologies used by the BEA give wildly different results. My point is that it is a waste of time: value the assets and liabilities using income – which is standard valuation practice.

      • Ramanan

        Will check.

        What about China’s stock market and Apple’s China unit?

        But those are more complicated things. I am yet to receive an answer about how infinite return of net income of $1 on net FDI stock of $0 is wrong.

      • Eric Lonergan

        Ramanan – why do u want to calculate a return on net assets? I don’t see the point. The point about Apple is to illustrate the nonsense of producing changes in the value of FDI using local stock markets. Have u looked at how market values of FDI are being estimated? If not you will keep misunderstanding my point.

  2. JKH


    Good post for information and analysis.

    But there’s no need to mangle the accounting in order to tell the story.

    The measurements are not “wrong”.

    For one thing, something you’ve overlooked is the overall US consolidated wealth effect.

    When the US NIIP is consolidated as a component contribution to US household net worth (as in the Fed Flow of Funds analysis), the equity value of those US firms making those foreign investments gets factored in on a net basis as either direct or indirect equity holdings of households. And that includes the market valuation effect of the effect of FDI investments held by those US firms, because equity issued by US firms is valued at market.

    That’s a consistent consolidation from the US perspective. Valuing US NIIP otherwise on a stand-alone basis would wreak havoc with that consolidation.

    Indeed, the US NIIP is operating analogous to a macro hedge fund with negative book equity and positive income. Not bad. But valuation of that hedge fund as a “stand-alone enterprise” is somewhat whimsical. On the supply side, it’s rather difficult to for the US to package it up for selling. On the demand side, there’s nobody on Mars or China that’s in a position to buy it.

    I say just tell the story behind the numbers.

    And for economists everywhere – please stop with the war on accounting.

    Accounting is not the enemy – be it NIIP, helicopter drops, or anything else.

    And BTW, risk analysis regarding the sustainability of this trend is an entirely different subject. Again, there’s no need to mangle accounting in order to do effective risk analysis.

    So I say just tell the story without reinventing the accounting wheel.

  3. Eric Lonergan

    JKH – no one even pretends that BEA net assets are proper accounting. Read the links. This is not corporate balance sheet accounting. Anyone who takes the numbers at face value, in this case, is not saying that standard accounting is correct, but that any estimate by a statistical agency is true even if it’s nonsense!

  4. Marko

    I’m as suspicious of official data as anyone , but once we agree to throw BEA , FOF , NIPA and assorted data collections overboard , it’s game over. You can then prove everything , and nothing.

    This recent paper , which I don’t see referenced above , touches all the valuation , privilege , and financial bases , and might add some useful perspective to this debate:


    • Eric Lonergan

      The point is that the income data allows you to value the net external position – you don’t need spurious ‘market values’. No one is throwing the data overboard. It’s very simple: value FDI using cashflows NOT proxies using stock market changes.

  5. Marko

    A DCF analysis makes sense only if you have some way to evaluate the stability of those income flows. A Trump presidency , for example , could easily blow a huge hole in the healthy yield differentials that now exist between U.S. and foreign assets.

    Your view might turn out to be correct over the medium-term , or even long-term , but nobody can say for sure. You can’t just arbitrarily slap a AAA rating on that income stream. The market-based “proxies” you disparage have some of that uncertainty built-in , and are probably better for policy analysis purposes. Let’s assemble some more convincing evidence in its favor before we attempt to massively import our way to prosperity.

    • Eric Lonergan

      Marko – if u read my two notes on this issue, I analyse the risk properties of income paid and received in both pieces. And US income received is less volatile. That’s a central point.

  6. JKH

    Agree generally with the comments of Ramanan and Marko.

    The BEA provides a suite of measurement alternatives for FDI, one of which is the way things are actually done in corporate accounting for FDI – book value – so that’s a first order fit in terms of overall measurement coherence. The alternative measures provide some additional colour in terms of replacement value and a crude type of proxy market value. I don’t think the latter measures are terribly useful, but they are less dangerous than what you propose.

    The NIIP is not a common stock for personal over the counter valuation. To slap an arbitrary 14 x multiple on this macro position in order to produce a more pleasant present value mapping to NIIP income is just accounting shenanigans. Ironically, but not atypically (among economists in general), this sort of accounting manipulation has nothing to do with economics. Neither Mars nor China is set to bid on this “market position.”

    As I said earlier, the value of FDI shows up mostly in the stock market value of US corporations, which is legitimately available information. The connection should be made there. In general, there are good, objective reasons why corporate balance sheet book value is not the same as stock market value, and this is just one of them.

    The point is that the balance sheet accounting alternatives for BEA are explained. Their inherent limitations are evident to those who understand the role of accounting. Invoking 14 times on the macro position is pointless in terms of understanding the economics of the position, which should be explained from first principles in terms of the cash flow patterns and the ultimate effect on legitimate objective measures of US wealth, such as the US stock market.

    • Eric Lonergan

      Hi JKH, You haven’t really addressed the substantive point. There is clearly an anomaly: the US has continued to generate a rising positive net income from its NIIP. Also, the income it receives is less volatile than the income it pays – over decades! At the same time the BEA estimates suggest rising and high net external liabilities.

      Those two facts are inconsistent.

      The BEA provides a range of valuation estimates, with very wide variance. Most people cite the ‘market value’ estimates. You agree that these are highly misleading. Are u really suggesting that the BEA’s book value estimates are likely to be better than PVs of income?

      The use of 14x is illustrative. The burden do proof is surely on those who want to value FDI going into the US on a higher rating than US FDI abroad, when the income flow is more volatile! My point is simply to put the income flows on the same rating.

      • Eric Lonergan

        JKH – to put this in context, a ‘market value’ estimate
        of foreign holdings of US treasuries is likely to be accurate. I think everyone would agree on that. But how do we value Google’s business in Europe? European stock market returns (which have been awful), are irrelevant – but that is what the BEA uses for ‘market value’ calculations. Are you really suggesting book value? This is the whole point: US overseas assets are generating far higher and less volatile income than foreign assets held in the US – that is not consistent with a large net liability. It really is as simple as that. Other issues, such as how the current account deficit frustrates domestic stimulus, or the vulnerabilities of the treasury bond market to capital flight (also nonsense, but for other reasons), are all totally separate matters.

        On the specific point: does the US have a large net external debt? The answer appears to be the opposite: it has large net external assets, and the reasons appear clear and persistent.

        As an aside, you seem to be suggesting earlier that we should use Google’s actual market cap to estimate the value of its overseas investments – that’s a great idea (some of the net income would be on more than 20x!). Only problem, no one is doing it – the BEA does the opposite: if the S&P500 goes up because of Google, Apple, and Facebook, the BEA writes up the value of overseas investments in the US!

  7. Ramanan

    Also agree with Marko.

    And it’s not just political risks. According to Eric’s logic, it doesn’t matter what current account deficit is. For simplicity assume the CAD is $2T every year in the next five years. Liabilities will be so large compared to assets that net income will turn negative even if return on assets is greater than liabilities.

    The DCF model assumes A&L aren’t changing in the future. DCF actually assumes CAB balanced in every period from now to eternity.

  8. JKH


    I think “that the US current account deficit is the accumulation of a debt that will one day need to be repaid” and “the US in fact has net external assets” are two different ideas, each of which is not necessarily correct – for different reasons. More particularly, the error in the first does not depend on the second being true. The first is wrong simply because *some* level of net external debt should be sustainable for the US, given its likely relative economic position in the world over the longer term, including the external demand for high quality assets such as US Treasuries and US currency. And this sustainability does not necessarily require a formal net asset position in total.

    In my view, that “the US in fact has net external assets” may be a debatable point, but it is a dubious claim based on the methodology you suggest, for reasons I’ve already given.

    But here’s one more: if the argument depends on the present valuing of net income in the form of more “realistic” FDI revaluations, I think that is very questionable. As I said, the value of FDI is embedded in parent stock market value. But that does not mean that FDI positions can be treated as stand-alone asset values using the same parent company valuation criteria. The value of FDI to the parent company is synergistic with parent company technology. That’s not the same as the value of FDI on a stand-alone basis. A buyer for such FDI could not be expected to generate the same income level in the absence of the original Google or Apple parent technology synergy for example. So I think it is incorrect to impute independent asset values to FDI on that basis. And that suggests a circumspect valuation for FDI as it appears in a NIIP representation. There is no reason to believe that FDI represents liquidation value in that position such that it could in theory or practice repay all liabilities such as US Treasuries and all the rest. The fact that the net position may still be sustainable when depicted as a net liability (as I have suggested) is beside the point in this context.

    So I still say – stick with the story in aggregate net income terms and be conservative in representing the net asset or net liability position. The existence of a nominal net liability such as is represented by FDI valued at book value should be no more threatening in concept that the existence of a book value for Google or Apple that is a fraction of its market value. It is the earnings stream that should provide comfort in the case of either these stocks or NIIP.

    • Eric Lonergan

      All good points JKH. I think we essentially agree – the idea that the US has some kind of external liability problem because of the current account deficit does not seem to stack up. I value assets and liabilities based on cash flows – so I would say the US has external assets. This is a simplification because the nature of the assets and liabilities are so different. That said, rising net income strongly suggests that the US is a significant net beneficiary of the exchange of assets, despite the fact that the overseas investment each year is lower than inward investment – the counterpart of the current account deficit.

      It should not surprise us when gross capital flows are so high, that the relative returns on capital dominate the effects of the current account.

      As a separate matter, I’m not a big believer in the use of book values for valuing most businesses. And we agree the BEA ‘market values’ for FDI make little sense.

  9. Ramanan


    I keep making the point on net return of $1 on a net stock of $0 to illustrate a point. That method isn’t a valid method to prove something.

    About stock indices, I think you are oversimplifying it. Since SNA/BPM prefers market values, it tries to estimate all values of their market value. It’s not like it is completely independent of what the company is and so on. It’s just what the company is worth in local markets and what the value for direct investor is.

    it’s a bit like house price evaluation using local markets. Nobody is saying the house price is independent of how it looks, what the garden is like, and other things unique to it. But prices of houses around this house matters. That way.

    • Ramanan

      I mean let’s say some US company has a stake in an Indian firm and it’s a direct investment. Won’t the value of assets for the non-resident firm be linked to the stock market in India?

      • Eric Lonergan

        Thanks Ramanan – take a look at the discussion below with JKH, & both my articles. It’s quite clear what method the BEA uses for FDI valuations – and why they are flawed. The key point is that if your investments generate superior returns, you can invest less and have more assets.

Leave a Reply

Your email address will not be published.

* Checkbox GDPR is required


I agree