On Natural Interest Rate Determination and Policy Rates
Reflections on Central Planning of Interest Rates
Summary
Are policy interest rates too high or too low? This study details why an answer may be an impossibility.
Centrally-planned interest rates may be incompatible with diverse firms generating a wide distribution of possible natural interest rates rooted in individual firm characteristics.
This Natural Rates series develops a micro-economic alternative to official estimates of the natural (or neutral) rate of interest to better visualize free market-based interest rates.
The analysis estimates natural rates of interest (NRIs) to fiscal year-end 2021 sourced from 110 datasets of global publicly-traded firms with an estimated market cap of over $15T.
Overview
This report updates an earlier version of this study (Kennedy 9/2021) by extending the underlying financial data to FYE 2021, adding 11 datasets, and removing redundant datasets. Analytical Refinements. A key feature of the study was to introduce refinements to the estimation of free market interest rates in two components. Since lenders typically wish to have their loans paid back with some level of certainty, the proposed refinement establishes a minimum level of creditworthiness upon which extension of credit in the loanable funds market would be based. Creditworthiness. The level of creditworthiness is specified by the MAE 0.10 estimate of net free cash flow to revenues (NFCF/Revenues) and corresponds to a theoretical maximum annual amount of financing obligations that the company can cover given the MAE specification of 0.10; (otherwise stated as the 10% of possible outcomes in which the company’s financing obligations cannot be covered). Note that any actual amount of financing obligations (defined as debt repayments and dividend payments) that exceeds this theoretical maximum amount would need to be restructured. Natural Rate of Return (abbreviated NRR) was previously an estimate of a central value (i.e., relative spatiotemporal centrality) of the company’s annual net free cash flows as a percentage of revenues (i.e., NFCF/revenues) defined as a maximum acceptable exposure MAE of 0.50. However, using a central point for cash flows is an unacceptably lax criterion for credit as it assumes a theoretical coverage capacity of only 50% of possible outcomes which does not meet minimum credit criteria for coverage. Therefore, the NRR is refined to correspond to a more stringent MAE 0.10 creditworthiness level as specified above.
For details on the datasets, other definitions, and assumptions, as well as a historical overview of natural interest, see the Appendix. Related previous natural review reports by the author are found in the References (Kennedy 2019a, b, c; 9/2021) and publications (Kennedy 2018, 2020, 2021).
Distribution of Natural Interest Rates
The following chart is the theoretical distribution of annual natural interest rates for all 110 corporate datasets. The x-axis represents firm-specific natural rates of interest, abbreviated NRIs; the vertical line indicates an arbitrary NRI of 5%.
Interpretation
In a free market for loanable funds, a single rate of interest is not compatible with the profiles of diverse businesses and many negotiated interest rates are expected to emerge. This may be seen by observing the proportion of estimated NRIs below and above a given value: The model suggests that 41% of the NRIs in the chart fall below 0%, and 59% of them lie above 0%, respectively. At an NRI of 5% (shown at the vertical line in the chart) the proportion of NRIs under/over 5% is roughly 78%/22%, respectively. For NRIs over 10%, the proportions of under/over 10% are 89%/11%, respectively (see the next table for the firms that fall into this top 11% category). While not conclusive, the distribution from which the estimates are derived ranks within the top one or two distributions in goodness-of-fit testing.
NRI<0%. At natural interest rates below zero percent there is no lending rate at which borrowers can cover their financing obligations. In such cases, the company must rely on investor financing, or some form of banking system forbearance in the form of “evergreening” or permanent working capital credit lines (such as the rolling over of existing obligations to repay obligations). As noted above, a non-trivial 41% of the NRIs are estimated to fall below 0%.
Elevated Natural Interest Rates (>5%). Higher natural interest rates correspond to companies that have strong net free cash flows (NFCF) at the MAE 0.10 level; these firms have stronger financing obligations coverage capacity (i.e., debt repayment and dividend payment) and can because of their strong net free cash flows, can not only afford to pay higher interest rates to secure credit, but can do so quite profitably.
Table. Which firms equal or exceed a 10% NRI? These firms and their estimates are shown in the table:
The columns are as follows: 1. The ticker symbol and Standard Industrial Classification Code abbreviated as SIC code; 2. Estimated unfiltered ranking of the firm according to the cautionary methodology; 3. MAE 0.10. The net cash flows as a percentage of revenues at the maximum acceptable exposure criterion of 0.10 which as detailed above is now redefined as the Natural Rate of Return NRR to specify the level of creditworthiness desired; 4. The estimated Natural Rate of Interest or NRI is a theoretical figure equal to half of the NRR and just like the NRR, varies across firms.
Qualification: Excessive Financing Obligations. The level of current financing obligations (abbreviated FO) of the firms cannot be ignored. As seen in the next table, PM, ABBV and BHP all are estimated to have excessive financing obligations (indicated in red font in the table). “Excessive” is computed as MAE 0.10 less financing obligations, both as a percentage of revenues. In words, this is the percentage of revenues that financing obligations exceed the theoretical minimal creditworthy level of MAE 0.10. For example, PM’s financing obligations of 43% of revenues exceeds its theoretical creditworthy MAE 0.10 level of 24% of revenues, therefore, the firm is estimated to have excessive FO of 19% of revenues.
Note that the ranking system is divided into unfiltered, and filtered rankings. Filtering may result in a downward adjustment of the ranking due to excessive financing obligations and other factors. Accordingly, PM’s filtered ranking drops from 1.6 to 4.0 (filtered ranking not shown in the table above). If a filtered ranking remains unchanged from the unfiltered ranking despite elevated financing obligations this may be because financing obligations (among other factors) did not reach a threshold that would trigger a reduction in the ranking.
APPENDIX
Datasets
The financial data are sourced from 10-K, 20-F filings or online databases. The sector composition of the datasets is shown in the following pie chart. Consumer goods constitutes the greatest proportion (30%) followed by healthcare/industrial (29%), and technology (23%). Total current market capitalization is estimated at more than $15T.
Notes. Dataset sample periods are generally small and limited to between five to ten years of either fiscal year end or interim financial data to the fiscal year-end 2021.
Sectors. Basic Materials includes SIC 1000 metals mining, 1040, gold & silver ores, SIC 1311 oil & gas production, and SIC 2911 petroleum refiners. Each dataset corresponds to the results for a single firm. This updated study removes the multiple datasets of KO that were included in Kennedy 9/2021 due to redundancies. NEC refers to not elsewhere classified.
Subsectors. There are twenty-five subsectors or categories within the sectors above, as follows:
1 Aerospace & Defense
2 Apparel and Accessories (2300-2389)
3 Auto Makers
4 Beverages
5 Chemicals
6 Computer Equipment-Diversified
7 Diversified Instruments
8 Dividend Stocks, not elsewhere classified (NEC)
9 Electrical Equipment & Components (3600-3629)
10 FAANG (tech) stocks, NEC
11 Food Products
12 Gold and Silver Ores
13 Household & Personal Products
14 Machinery/Equipment Makers (non-electrical)
15 Metals Mining
16 Oil & Gas Production
17 Petroleum Refiners
18 Pharmaceuticals
19 Retail
20 Semiconductors
21 Software/Programming
22 Telecommunications
23 Tobacco
24 Utilities (Electricity 4931)
25 Utilities (Water 4941)
Summary of Assumptions and Definitions
Natural interest rates constitute a starting point to help identify where in the absence of a centrally-controlled market, where free market rates may find their equilibria in free markets for loanable funds; interest rates range across the spectrum and credit is allocated according to the range of returns (interest rates) for lenders.
Natural Rate of Return NRR is annual net free cash flow divided by revenues (NFCF/Revenues) estimate at the MAE 0.10 specification (see “credit requirement” below). This return is firm-specific and is also the firm’s return on its own cost, relating their net free cash (in)flows to their costs (outflows) synchronously. Note: In this update, the NFCF/Revenues variable is used as a proxy for NFCF/Cost because revenues are broadly used as the common denominator, and although there is a difference between the two ratios, the NFCF/Revenues produces a more conservative (lower) NRR.
Natural Rate of Interest (NRI): This rate is assumed to be one-half of NRR and is also firm-specific, see above. In this study, the interest rate is annual, for instance, on a generic working capital credit line that can be cleaned up and fully repaid. NRI may also be labelled a natural cost of funds.
Credit Requirement. Prior to this study, the MAE specified was 0.50 (i.e., relative spatiotemporal centrality) which assumes an essentially indifferent credit requirement in which coverage of financing obligations is only assured in half of the possible outcomes. This study introduced a key refinement of the previous assumption of MAE 0.50, basing the extension of credit on a more stringent credit guideline of maximum acceptable exposure MAE of 0.10. A coverage ratio is not applicable in this methodology because the credit requirement of MAE 0.10 defines the coverage capacity: Financing obligations coverage is in theory achieved in 90% of possible outcomes. This considerably simplifies the analysis of debt service coverage.
Natural Rates and Market Distortion: Review of Components
We begin with a quick review of terminology: As defined above, natural rates of return and interest are firm-specific and are as diverse as the number of firms in existence. Market rates of interest emerge from two kinds of systems: a) in free and undistorted loanable funds markets they are negotiated interest rates based on supply and demand for funds without external monetary influences (i.e., free market interest rates) and b) policy-influenced interest rates which vary too but tend to be controlled to remain with an acceptable established band of desired rates, and to conform with political considerations and corporate special-interest pressures.
In both systems, market rates of interest become the cost of funds (abbreviated COF for firms; one is a free market cost of funds, the other is a policy-influenced cost of funds).
In a centrally-planned monetary system, market interest rates may “split” into two distinct sub-elements that are linked by the monetary policies that help generate the split: (a) the policy-influenced COF according to debt maturity and (b) the interest rate gap/distortion (representing a distortion influenced by policy); it is thought that this “split” can be roughly detected by the difference (or “gap”) between a given firm’s natural rate of interest NRI and the policy-influenced COF for the market. A chronically positive interest rate gap regardless of market conditions suggests that the current cost of funds is being maintained low by policy (i.e., an artificially wide spread), allowing borrowers to continue benefiting from a greater spread on their borrowings than otherwise possible (this also facilitates a wealth transfer from savers and lenders to these borrowers). Example of the breakdown in a distorted market: Let’s assume that the natural rate of return (NRR) of a firm is 10%, the firm-specific NRI is half of this or 5% by the assumption above. Borrowing Profit: This difference between the NRR and the NRI for the firm is akin to the theoretical “profit” to this firm of borrowing funds. This hypothetical 5% NRI defines a rough ceiling desired by the borrower to lock in that profit from borrowing and exerts some influence on the free market depending on the size of the loan request, etc., but overall supply and demand of loanable funds as well as the (high or low) NRRs of competing firms will eventually establish free-market rates negotiated between borrowers and lenders.
In contrast, in a controlled environment, the policy-influenced cost of funds (for the same debt maturity) may be held much lower than 5%--in this example it is assumed to be cut to 2%; this implies that the interest rate gap/distortion is 3% (5%-2%=3%). This 3% represents a distortion, meaning that the cost of funds is being artificially kept 3% lower than otherwise; highlighted in yellow in the diagram below. The following simplistic diagram is constructed to convey the scenario and the differing systems:
Self-Regulating Dynamics of Free Markets. An example of rate movements in undistorted market conditions is briefly explained: Supply and Demand. The cost of funds COF is expected to rise when there is a glut of requests for borrowings relative to the supply of loanable funds, or from individual firm NRRs (potential borrowers with higher NRRs enter the market and are able to profitably borrow at higher cost of funds rates than the others, therefore outcompeting other firms for borrowed funds; a rising cost of funds would automatically be an automatic disincentive for competing firms to borrow more funds unless they can increase their NRRs through various economies/cost reductions, etc. This dynamic induces self-regulating credit rationing. Put another way, for a firm with a given NRR, when the cost of funds rises in the market for loanable funds, the firm’s nominal borrowing spread shrinks, making it less profitable for that firm to borrow. This process continues whereby borrowing naturally slows down when it is less profitable for firms to do so, and in response lenders may choose to accept a lower return (interest rate) on new extensions of credit, ad infinitum.
Note: A free market rate of interest is not known in advance and therefore can only be approximated based on individual firm estimated NRRs and NRIs, as well as loanable funds market supply and demand conditions. In a policy-influenced market system, since NRRs (and NRIs derived from the NRRs) vary by firm, there can be some cases where the policy-influenced COF rate happens to closely match a firm’s natural rate of interest; it is not clear whether this is coincidental or an intent of policy to accommodate a specific firm’s NRI. This rough matching phenomenon was observed for Ford Motors (F) and Exxon Mobil (XOM) in Kennedy 2019b where the interest rate gap/distortion was found to be near zero because their estimated 10-year NRIs of 2.5% closely approximated the then policy-influenced 10-year cost of funds of 2.5%; considering that these two large firms with heavy borrowing needs may exert an outsized influence on markets and may also have some political sway as well, it is possible that some form of policy-influenced accommodation may play a role.
Rankings and Natural Interest Rates: A Comparison (Chart)
The following chart from the earlier version of this study shows visually how estimated rankings of firms using the cautionary methodology correspond to the natural rates of interest estimated from the datasets.
The ranking system consists of 4 tiers: The top three tiers are 1 through 3 with a range of 1.0 to 3.99; the 4.0 tier is a catch-all (or remainder) category for companies that did not meet the criteria for the top three tiers; there is no ranking within the 4.0 tier, with one exception--a single special category called 4.0/unranked which is designated for firms that are viewed as fundamentally infrastructural due to high capital expenditures and/or acquisitions activity with heavily negative cash flow profiles; note that this category is digitized here as “4.5” for purposes of plotting on the chart.
Chart. As a result, the plotted datapoints on the chart reflects two paradigms: a) granularity for the ranked tiers 1 through 3 (with a concentration found in the high threes) and b) the largely unranked tier 4.0; this may be visible in the chart in a tendency towards flattening and outliers beginning at approximately 4.0; accordingly, the regression analysis detailed further below excludes values of 4.0 and greater. The natural interest rate NRI is on the horizontal axis and the rankings on the vertical axis.
Linear Regression. For reasons stated in the previous paragraph, rankings of 4.0 and above are excluded from the sample, leaving 61 observations for estimation by ordinary least squares OLS. The relationship between the rankings (i.e., the top three tiers, rankings 1.0 through 3.99) and the natural rates of interest NRI suggest a significant negative relationship as follows:
R= 3.27 - 8.63nri; (40.5) (-6.96); Radj=0.44; F=48.4.
This suggests that as the rank declines (meaning stronger fundamentals), the natural rate of interest rises. The variable nri is the natural rate of interest and the dependent variable R is the ranking. Figures in parentheses are the t-values that correspond to the estimated coefficients of the equation. Radj refers to the coefficient of determination (adjusted R-squared value) and “F” is the F-statistic. Note that the shortcomings of regression analysis are recognized, and a simplistic assumption of linearity may not be valid.
Historical Overview
Modern Approaches. The natural or neutral rate of interest as modelled by central banks rely largely on macroeconomic variables; Knut Wicksell (1898), with whom the term ‘natural’ interest rates may have originated, is noted in the references. Wicksell’s role was important as he was clearly familiar with a classical school of interest theory detailed below and was able to provide a market-based guideline for central banking in setting policy rates (re: Brainard 2015; Holston, Laubach and Williams 2016, Manrique and Marqués 2004, Salerno 2016; Spitznagel 2017).
Also marking a critical shift to the modern view is the work of Irving Fisher (1907, 1930) that redefined the classical term ‘real interest rates’ to mean something entirely different, and which remains predominant in modern interest rate theory (re: Fisher equation; Fabozzi and Modigliani, 1992).
Classical View. The classical literature regarding interest, which dates at least as far back as John Locke (1668), is associated with a view of natural interest rates reflecting profitability, rate of profits or expected rates of return; these business rates of return are generated in the process of productive economic activity including capital investment. Crucial to note is that the natural rates of interest derived from the rates of return of firms are sharply distinguished from the interest rates offered through the banking system (i.e., bank loans as inducements to borrow which were criticized for distorting economic activity through malinvestment). Also refer to Cantillon (1755), Smith (1776), Ricardo (1826), Bohm-Bawerk (1901), Mises (1912), Fetter (1914), Ropke (1936), among others. Terminology can be confusing because different words describe similar concepts: The term natural interest (Wicksell), real interest rates, normal interest rates, market interest rates all had meanings centering around this classical concept. Keynes (1936) added to the variety by coining the term marginal efficiency of capital. Redefinition, Critically, Fisher, noted above, used the term ‘real interest rate’ but altered the meaning of ‘real’ away from the classical sense by defining ‘real’ as equal to ‘nominal’ less inflation in his eponymous Fisher equation. An approach to defining the classicists’ rates of return is detailed next.
Natural Returns: Framework and Methodology. A shortcoming of the classicists may be their limited deconstruction of the financial dynamics of firms to define rates of return. Bridging this gap is a primary aim of this microeconomic-financial approach and previous studies by the author listed in the References. The term ‘natural’ is retained; the natural rate of return (NRR) refers to a measure that captures some fundamentally inherent value for the entity's cash rate of return on its cost (in this study updated to replace cost with revenues*). Cash, rather than accrual basis, is fundamental because rates of return are ultimately accounted for physically (i.e., in terms of physical cash inflows and outflows, including existing debt service and dividend payments).
NRR. In a context of a financial metrics nomenclature, these natural rates of return (NRR) have been termed rates of return on cost (RRC); both NRR and RRC are the same. RRC is defined as the firm’s net free cash flow (NFCF) divided by Cost*; both are annual figures in the same fiscal year, cash basis. NFCF is free cash flow (FCF) less acquisitions and related activity (also referred to by the author as equity income); cost is the total of the cash cost of sales, S, G &A, capital expenditures and acquisitions related activity.
*Note: As detailed above, in this study NFCF/Revenues is used as a proxy for NFCF/Cost.
Credit Guideline. A major refinement to the natural rate of return NRR estimation introduced here is establishing a more stringent criterion for credit extension defined by a maximum acceptable exposure MAE of 0.10.
This work is shared as a public service; if helpful, consider paying it forward by adding something extra to any donations made to reputable charities, preferably with priority given to the most vulnerable, including defenseless animals. Organizational reputations may be researched through sites such as charitynavigator.org.
The author may also hold positions in securities of companies, including through ETFs, which are covered herein. The discussion and any visuals may contain significant errors, are subject to revisions and are provided 'as is' solely for informational purposes, not for trading or investment advice. This preliminary analysis is exploratory; no claims are made as to the validity of data, assumptions, theoretical models, and methodologies; results may be based on prior data that do not reflect the most current market or other events.
References
Bank for International Settlements (BIS), "Global Financial Markets Remain Dependent on Central Banks," Chapter II, BIS 85th Annual Report 2014/15, June 28, 2015, pp. 25-43.
Bohm-Bawerk, E. von., “The Function of Saving” Annals of the American Academy of Political and Social Science (May 1901).
Borio, C. and Disyatat, P.., " Low Interest Rates and Secular Stagnation: Is Debt a Missing Link?" VOX, June 2014.
Brainard, Lael., "Normalizing Monetary Policy When the Neutral Interest Rate Is Low," Presentation at the Stanford Institute for Economic Policy Research, Stanford, California, December 1, 2015.
Cantillon, Richard., Essai sur la Nature du Commerce en Général (traduit de l’anglais), 1755.
Clarida, Richard., “Removing Accommodation” PIMCO Global Central Bank Focus, March 2017.
De La Calle, L.S., (1544) Instrucción de Mercaderes ... (see Grice-Hutchinson 1952) (Re: School of Salamanca)
El-Erian, Mohamed., The Only Game in Town: Central Banks, Instability and Avoiding the Next Collapse, Random House, New York, 2016.
Fabozzi, Frank J., The Handbook of Fixed Income Securities, 8th Edition, McGraw-Hill, 2011.
Fetter, Frank A., "Interest Theories, Old and New," The American Economic Review, Volume IV, No. 1 (March 1914), 68-92.
Fisher, I., The Theory of Interest (as Determined by Impatience to Spend Income and Opportunity to Invest It), New York: The Macmillan Company, 1930.
Grant, James., Mr. Market Miscalculates: The Bubble Years and Beyond, Axios Press, 2008.
Grice-Hutchinson, M. (1952)., The School of Salamanca: Readings in Spanish Monetary Theory, 1544-1605, 1952. (Also see Hutchison, Terence)
Hayek, Friedrich A., (1974), “The Pretense of Knowledge,” essay presented at the occasion of Hayek’s Nobel Prize in Economics, December 11, 1974, reprinted with permission by the Foundation for Economic Education (FEE).
Holston, Kathryn., Laubach, Thomas., and Williams, John C., “Measures of the Natural Rate of Interest: International Trends and Determinants,” Federal Reserve Bank of San Francisco Working Paper Series, Working Paper 2016-11, December 15, 2016.
Huerta de Soto, Jesus., Money, Bank Credit, and Economic Cycles, Ludwig von Mises Institute, 2012.
Hutchison, Terence., Before Adam Smith: The Emergence of Political Economy, 1662-1776, Basil Blackwell, Ltd., 1988.
Juselius, M., Borio, C., Disyatat, P., and Drehmann, M., "Monetary Policy, the Financial Cycle and Ultra-Low Interest Rates," International Journal of Central Banking, (13:3), 2017, pp 55-90.
Kennedy, Raoul., Financial Distortion Dynamics, 2018.
Kennedy, Raoul (9/2021)., “Natural Rates Review: What’s The Right Tweak for Policy Rates 3,” Seeking Alpha, September 2021. https://seekingalpha.com/instablog/471740-raoul-kennedy/5642214-natural-rates-review-what-right-tweak-for-policy-rates-part-3
Kennedy, Raoul (2019c)., “Natural Rates Review: What’s The Right Tweak for Policy Rates 2,” Seeking Alpha, November 22, 2019. https://seekingalpha.com/instablog/471740-raoul-kennedy/5377700-natural-rates-review-what-s-right-tweak-for-policy-rates-2
Kennedy, Raoul (2019b)., “Natural Rates Review: What's The Right Tweak for Policy Rates?” Seeking Alpha, May 6, 2019. https://seekingalpha.com/instablog/471740-raoul-kennedy/5301000-natural-rates-review-right-tweak-for-policy-rates
Kennedy, Raoul (2019a)., “Natural Rates Review: Intro,” Seeking Alpha, May 6, 2019.
https://seekingalpha.com/instablog/471740-raoul-kennedy/5300975-natural-rates-review-intro
Kennedy, Raoul., Yield Distortion Dynamics, 2020.
Kennedy, Raoul., Dividend Coverage Analytics, 2021.
Kennedy, Raoul., Debt Management Analytics, 2021.
Laubach, Thomas, Williams., John C., “Measuring the Natural Rate of Interest Redux” Federal Reserve Bank of San Francisco Working Paper Series 2015-16, October 2015.
Laubach, Thomas, Williams, John C., “Measuring the Natural Rate of Interest,” Review of Economics and Statistics (Vol 85, No. 4: 1063-1070) November 2003.
Locke, John (1668) as cited in Hutchison (1988: 63). See Hutchison, Terence.
Lundvall, Henrik., and Westermark, Andreas, “What is the Natural Interest Rate?” Sveriges Riksbank Economic Review (2011:2).
Manrique, Marta., Marqués, José Manuel., “An Empirical Approximation of the Natural Rate of Interest and Potential Growth” Documento de Trabajo No. 0416, Banco de España, Madrid, 2004.
McKinsey Global Institute, "Debt and (Not Much) Deleveraging," MGI Report, McKinsey & Company, February 2015.
Mill, John Stuart (1848)., Principles of Political Economy, Books IV and V, Penguin Group, 1988. (Original 1848)
Mises, Ludwig von., (1912) The Theory of Money and Credit (translated from the German) 1934; also, Yale University Press, 1953.
Ricardo, David (1826)., (P. Sraffa, Ed. with M. Dobb)., The Works and Correspondence of David Ricardo, Volume I: On the Principles of Political Economy and Taxation, Cambridge: Cambridge University Press, 1951-1983.
Ropke, Wilhelm., Crises and Cycles (adapted and revised by Vera C. Smith), William Hodge and Co. Ltd., 1936.
Salerno, Joseph T., “The Fed and Bernanke Are Wrong About the Natural Interest Rate,” Mises Wire, June 26, 2016.
Sigurjónsson, Frosti., Monetary Reform: A Better Monetary System for Iceland, Reykjavik, Iceland, March 2015. (Report commissioned by the Prime Minister of Iceland, Edition 1.0, 2015).
Smith, Adam (1776), An Inquiry into the Nature and Causes of the Wealth of Nations, David Campbell Publishers Ltd., 1991 and University of Chicago Press, 1976.
Spitznagel, Mark., "What Is This "Neutral" Interest Rate Touted by the Fed?", Mises Wire, January 3, 2017.
Stockman, David., The Great Deformation: The Corruption of Capitalism in America, PublicAffairsTM, 2013.
Taleb, Nassim Nicholas, et al., Statistical Consequences of Fat Tails: Real World Preasymptotics, Epistemology, and Applications, STEM Academic Press, 2020.
White, Lawrence., “Antifragile Banking and Monetary Systems,” Cato Journal, Vol. 33, No. 3 (Fall 2013), 471-484.
Wicksell, Knut., Geldzins und Guterpreise, (Interest and Prices) Jena, 1898 (English-language translations 1936, 1965: Interest and Prices, New York, Kelley 1965).
A brief video summary is available at https://youtu.be/DhBmjX0wF3Q