Liquidity, Inequality, Charity and Temerity

32 min readDec 28, 2022

(Originally published on LinkedIn on December 28, 2016)

Overview

In recent years, it has become commonplace to argue that expansionary monetary policy has flooded liquidity into the markets, driving up the prices of all kinds of risky assets. One widely reviled set of “culprits” for this behavior are charitable foundations and endowments, who are well known for favoring asset classes such as venture capital and hedge funds. The mechanisms through which endowments soak up the liquidity that comes from expansionary monetary policy are, however, not really understood. Nor is there much clarity about what propels endowments to make such risky investments.

In this paper, we seek an explanation for how monetary expansion induces endowments to make risky investments. We argue that monetary expansion (“Liquidity”) worsens the distribution of income, consumption and wealth (“Inequality”). This happens through five channels:

(1) Income Composition Channel: profits gaining at the expense of wages

(2) Financial Segmentation Channel: those closest to the financial system gaining at the expense of those further away

(3) Borrowers versus Savers Redistribution Channel: asset-holders gaining through interest rate cuts while borrowers find their liabilities revalued upward, the latter particularly acute for hand-to-mouth borrowers

(4) Earnings Heterogeneity Channel: while monetary stimulus is generally favorable for labor employment, income and consumption, those at the upper-end of incomes could benefit disproportionately if this happens under a less than progressive tax and unemployment benefits system

(5) Portfolio Channel: Expanding the supply of money reduces its return, hurting the poor who tend to largely hold their little wealth in the form of cash even as the purchase of both safe and risky assets enabled by the expanded monetary base benefits rich asset owners.

With increased Inequality from Liquidity via these five channels, there is a clear impact on the overall level of charitable donations (“Charity”). High income households that have high propensity for charitable action find more resources available to do so and also tend to care less if other households do not contribute their share. They find it possible to contribute even more if there is more demand for the charitable activities they support e.g. when times are hard. The empirical evidence seems to suggest that wealthy donors increase their charitable donations with higher income and wealth. But charitable donations depend not only on redistribution effects experienced say through increased equity or real estate prices but also on the demand for donations from endowments themselves. Charitable endowments are found to actively court donors when need or demand for their services is high; indeed, they also tend to themselves hoard more at such times. As an outcome of all of the above, endowments keep receiving regular donations through good times as well as bad; however, they do see some ups and downs in line with equity / real estate fluctuations. This pattern of flows can easily be recognized as similar to an annuity income with an upside for equity and real-estate prices; in other words, some kind of “synthetic bond with an equity kicker”.

From the above, we see that Liquidity leads to Inequality which in turn encourages Charity. The “synthetic bond with equity kicker” nature of the charitable donations flow has its own implications. To balance out the low-risk nature of the “synthetic bond with equity kicker”, holders of the bond must take more active risk. They do this through investing in high-risk asset classes such as venture capital and hedge funds (“Temerity”).

In summary, Liquidity exacerbates Inequality which encourages Charity and in turn necessitates Temerity. We hypothesize that this is how expansionary monetary policy leads charitable foundations and endowments to invest in risky asset classes.

Liquidity Exacerbates Inequality

As macroeconomic theory evolves and as empirical work grows, there is increasing realization that monetary policy has redistributive impacts on income, wealth and consumption.

Balac (2008)’s study reviewing trends from the early 1960s to early 2000s shows that monetary expansion seems to affect measures that emphasize the extremities of the income distribution (Theil, Atkinson) more than those that tend to weight median income (Gini); this leads the author to worry about monetary expansion leading to “coercive redistribution of wealth”.

There have also been a number of studies covering the Great Recession period. Watkins (2014) shows that while all deciles of US households suffered a decline in net worth between 2007–10, the net worth share of the top decile suddenly jumped from 81% to 86%. Applying the evolution of asset prices and assuming portfolio compositions to be relatively stable, Domanski Scatigna Zabai (2016) simulate changes to initial household balance sheets after the global financial crisis (in the US and five other developed economies). During the period covered by their simulation, the net wealth of richer households grew four times as fast as that of poorer ones in the US. If their simulation accurately represents underlying trends, after 2010, high equity returns have been the main driver of faster growth in net wealth at the top of the distribution. The impact of equity prices on inequality seems to be much more cyclical, and less persistent, than that of house prices, which exhibit long booms and busts. The direct impact of low interest rates and of rising bond prices on wealth inequality may have been small but differences in household leverage have amplified distributional effects. This is because when households are highly leveraged, asset gains or losses have a larger impact on wealth and these impacts accumulate over time. Meanwhile, Saiki and Frost (2014)’s VAR study of two periods of Japanese quantitative easing, 2002–2006 and 2008–2013 shows that the monetary base has a statistically significant impact on the Gini coefficient and that the more aggressive second phase of easing also had sharper impact on inequality.

To be sure, a few researchers do dispute that asset purchases by central banks increased inequality. Bivens (2015), for example, points out that prices of homes — the key asset of poorer households — have also risen together with equity prices. However, the more widely prevailing view is that central bank asset purchases have contributed to inequality.

While the circumstantial and statistical evidence is highly suggestive of monetary stimulus having redistributive impacts, the underlying causal mechanisms are also becoming better known. Coibion et al. (2012) summarize five channels that have been referred to in formal or popular literature whereby monetary policy can impact income inequality. Interestingly, Coibion at al themselves seemed to believe that the net outcome of the five channels was that contractionary policy increased inequality while expansionary policy reduced inequality. However, subsequent literature seems to generally agree that monetary expansion mostly exacerbates inequality and even Coibion et al in a later note pointed out that negative rates could actually increase inequality.

Income Composition channel:

The income composition channel refers to the shares of national income received by different factors of production e.g. capital vs labor. Coibion et al (2012) point out that if expansionary monetary policy shocks raise profits more than wages, then those with claims to ownership of firms will tend to benefit disproportionately. Since the latter also tend to be wealthier, this channel should lead to higher inequality in response to monetary policy shocks.

As it happens, US data do indeed show that rising income inequality after 1967 was mostly due to an increase in non-wage incomes — i.e. business income, capital income and capital gains. (The Income Composition Channel sounds superficially similar to the Financial Segmentation Channel and also the Earnings Heterogeneity Channel but differs from both as we will see below).

Bezemer (2014) explains the mechanism that caused this rising share of profits relative to wages. Data shows that increase in the GDP share of US total bank credit over 1952–2012 was mostly driven by growth in credit to asset markets (except after the 2007 financial crisis, as GDP contracted). This flow was emitted from banks and absorbed up front by the non-bank financial sector, not directly entering the real sector or directly financing physical investment or wages. Instead it was going to property and financial assets connected to pension funds, savings institutions, credit unions, funding corporations, mortgage pools, exchange traded finds, private pension funds and money market mutual funds. Ballooning financial and insurance markets and the housing bubble, as well as the plethora of derivative products developed in these markets provided ample investment opportunities to realize capital gains and capital incomes.

In summary, monetary expansion triggered the growth of credit to asset markets and financial instruments rather than investment or wages. This increased the share of profits relative to wages.

Financial Segmentation Channel:

A second channel that Coibion et al (2012) have found to have been discussed in the literature is proximity to the central bank and financial markets. An increase in the money supply will redistribute wealth toward those agents most connected to financial markets. While this may sound similar to the Income Composition Channel, this is a slightly different point: it relates not to shares of profits and wages but rather to those connected to the financial system versus those who are not.

Andrei and Ledoit (2015) point out that central banks may not be able to inject money uniformly (i.e., the “Friedman helicopter drop”) for either institutional or practical reasons. Instead, only a few select institutions have direct dealings with the central bank, while general members of the public must go through a chain of intermediaries. The money injected will eventually percolate through the economy, but the agents who interact directly with the central bank — such as government and investment banks — will beneﬁt from monetary expansion. In contrast, the agents with the furthest economic links — “Main Street” — will suﬀer from monetary expansion.

A model developed by Williamson (2005), where some households receive a money injection by the central bank while others do not, triggers price dispersion across markets and thereby consumption risk. Money growth shocks accordingly have small effects on aggregate employment but large effects on the dispersion in consumption.

Cohan (2014) expressed the popular version of the same sentiment when he raged: “Since the onset of Mr. Bernanke and Ms. Yellen’s policy, the Fed’s balance sheet has grown to $4.5 trillion, from around $800 billion before the crisis. That’s a whole lot of securities bought at high, profitable prices and paid directly to Wall Street traders. The Fed might as well have been paying the traders sevenfigure bonuses directly” or “If you’re General Electric or Kohlberg Kravis Roberts, getting a loan from a bank is no problem; if you want to buy a new house in Peoria, good luck to you”.

In summary, monetary expansion benefits those most closely linked to the central bank at the expense of those further away.

Borrowers versus Savers Redistribution Channel:

Coibion et al (2012) summarize the third channel thus: an unexpected increase in interest rates or decrease in inflation will benefit savers and hurt borrowers, thereby generating an increase in consumption inequality to the extent that savers are generally wealthier than borrowers.

Auclert (2016) refers to earlier empirical studies that show that net nominal household balance sheet positions are very positive for rich and old households but negative for the young middle class with ﬁxed-rate mortgage debt. On the surface, this implies that young, middle class indebted households would be beneficiaries of lower interest rates that come with expansionary monetary policy. In his own work, however, Auclert (2016) probes the issue in greater detail. It turns out that it is not merely interest flows or for that matter straight revaluation of assets and liabilities that matter. Instead, it is unhedged interest rate exposure that is the relevant issue. Unhedged interest rate exposure is the difference between all maturing assets (including income) and liabilities (including planned consumption). Such understanding of the composition of the household’s balance sheet is important to understand consumption, labor supply, and welfare response to changes in interest rates and prices. Because unhedged interest rate exposure includes the stocks of ﬁnancial assets rather than interest ﬂows, it can signiﬁcantly diverge from traditional measures of savings, in particular if investment plans have very short durations.

Applying this framework, Auclert (2016)’s analysis shows that monetary policy affects consumption activity via redistribution in two ways. The first is an asymmetric effect that can be traced back to the behavior of those households that are at their borrowing limit. While these households have to cut consumption one for one in response to increased interest rates, their marginal propensity to consume out of moderate interest rate cuts is low, Auclert’s (2016) estimate is below 0.3. A second effect is driven by the behavior of wealthy individuals holding longer maturity assets. Under longer asset maturities, expansionary monetary policy creates more capital gains for asset holders and additional upward revaluation of liabilities for borrowers. These capital gains and losses redistribute against the economy’s marginal propensity to consume gradient; or in other words, lower real interest rates tend to beneﬁt asset holders with lower marginal propensities to consume.

Accordingly, with the more detailed understanding of the Borrowers / Savers channel that Auclert (2016) provides, expansionary monetary policy is found not symmetrically opposite to contractionary policy. Lower interest rates have only a muted effect on stimulating consumption — because they are generally insufficient to help leveraged households but also because they benefit long-maturity asset holders with low propensity to consume.

In summary, monetary expansion exacerbates inequality by benefiting long-maturity asset owners while not providing much relief to indebted households.

Earnings Heterogeneity Channel:

Coibion et al (2012) summarize the fourth channel in the following manner. Labor earnings are the primary source of income for most households and these earnings may respond differently for high-income and low-income households to monetary policy shocks. While this may sound similar to the Income Composition Channel, this is a slightly different point: it relates not to shares of profits versus wages but rather to how wages are distributed among households.

In contrast to other channels, the Earnings Heterogeneity channel is one where monetary expansion appears to act to reduce inequality through stimulating employment. Indeed, the analysis of Gornemann et al (2016) suggests that wealthy households and median-wealth households have different preferences with regards to monetary stimulus. The wealthiest prefer a smaller monetary stimulus than if all households were homogenous. The median-wealth household, instead, favors a monetary stimulus to unemployment that is about twice as strong as the optimal response in a model where all households are homogenous. Of course, it is these benefits of stimulating employment and consumption that lead proponents such as former Fed Chairman Bernanke to aggressively espouse expansionary monetary policy.

The reasons for this are not hard to understand. “Labor earnings are the primary source of income for most households and these earnings may respond differently for high-income and low-income households to monetary policy shocks. This could occur, for example, if unemployment disproportionately falls upon low income groups. Similar effects could arise even for the employed in the presence of different rates of wage rigidities across the income distribution (e.g. from unionization in production but not management), varying degrees of complementarity/ substitutability with physical capital depending on agents’ skill sets (since interest rates affect the relative price of capital and labor), or different endogenous labor supply responses reflecting specific household characteristics such as age and number of children which may systematically differ across the distribution”. Heathcote, Perri and Violante (2009) looked at micro-data and found that business cycles are more dramatic for household earnings at lower percentiles of the income distribution. Wages don’t fluctuate as much as household earnings, clearly pointing to unemployment as the underlying cause. Accordingly it is no surprise that earnings inequality widens sharply at times of recession. Interestingly, another study shows that fall in consumption in recessionary times is even greater than actual falls in earnings would suggest. The authors of this study explain: “faced with the elevated chance of becoming unemployed, impatient households who have not yet lost their jobs and have currently medium to high income realizations start to save more for precautionary reasons”.

To combat the hit to consumption in recessionary times, less wealthy households want to see an improvement in their labor income. If expansionary monetary policy can do the trick, they would of course favor such policy. In Kaplan Moll and Violante (2016)’s Heterogeneous Agent New Keynesian framework, monetary policy affects aggregate consumption primarily through indirect eﬀects that arise from a general equilibrium increase in labor demand. Their model closely follows real-world household portfolios, wealth distribution and consumption behavior. Hand-to-mouth households are highly responsive to labor income changes and unresponsive to interest rate changes. And, although Kaplan Moll and Violante do not refer to it, the same effects are also likely to additionally be in play for households concerned about job loss as per Krueger Mitman Perri’s (2016) analysis. In any case, Kaplan Moll and Violante argue, the monetary authority must rely on powerful indirect general equilibrium feedbacks that boost household labor income by increasing aggregate demand in order to induce an economic expansion.

The stimulative effect of expansionary monetary policy on employment and consumption is widely agreed upon based on empirical evidence. However, it is not the whole story with respect to inequality. Auclert (2016) shows that fiscal policy which impacts earnings risk over the cycle — specifically the progressiveness of taxation and the extent of automatic stabilizers such as unemployment benefits — modifies the expansionary effect of monetary policy. Under a less-progressive tax and benefits system, monetary expansion could in principle disproportionately favor the highest-income individuals. The author defines a certain factor “gamma”, representing the cyclicality of earnings risk based on fiscal transfers, which determines the magnitude and sign of the Earnings Heterogeneity effect. Only if gamma is negative is consumption inequality reduced. Otherwise, monetary stimulus has either no effect on inequality via the Earnings Heterogeneity channel (gamma = 0) or can even worsen inequality along with other channels (gamma > 0).

In summary, monetary expansion generally reduces inequality when it works through the Earnings Heterogeneity. However, the actual magnitude of inequality reduction depends on the net progressiveness of taxes and benefits. If not so progressive, the reduction in inequality is weak and in the extreme case can even turn into an increase in inequality.

Portfolio Channel:

The Portfolio Channel is the last of the five channels identified by Coibion et al. (2012) who summarize it as follows: since poorer households hold relatively more cash in their portfolio, inflationary central bank actions represent a transfer from low-income households towards high-income households which tends to increase consumption inequality. Apart from the Heterogeneous Cash Holdings mechanism, however, the literature also identifies a second mechanism — Asset Prices — through which the portfolio channel operates. Both mechanisms are detailed below.

Heterogeneous Cash Holdings: Albanesi’s (2007) model with heterogeneous households shows that poorer households hold more cash for two reasons. The first relates to heterogeneity in transaction patterns which arises endogenously from differences in labor productivity. Low skill agents hold more cash as a fraction of total purchases. The second relates to idiosyncratic shocks which cause precautionary asset accumulation. Even though it may be suboptimal at social level, households may hold cash as a store of value, either through experiencing a history of positive shocks or as self-insurance against negative shocks. Empirical evidence in the US and other countries also strongly suggest a negative correlation between labor income (and wealth) and cash holdings.

Both Albanesi (2007) and an earlier paper by Bhattacharya, Haslag and Martin (2004) show how this heterogeneity in cash holdings changes the impact monetary policy has on inequality and welfare. The Friedman Rule states that “since money is an asset, the central bank ought to change the stock of outstanding money at a rate that causes the real rate of return on money to equal the real return rate on other physical assets”. Put another way “any positive value of the nominal interest rate effectively amounts to a distorting tax on cash holdings” and, indeed, the Friedman Rule “requires that inflation is equal to the negative of the real interest rate”. However, in the face of heterogeneity of cash holdings among households, the Friedman Rule breaks down and instead triggers redistributive effects. An expansion in money to move towards zero nominal rate not only lowers the rate of return on money but it also creates a change in overall price level. Under equal holdings assumptions implicit in the original Friedman Rule, this would have no impact. However, with heterogeneous holdings, money supply expansion has differential impact on households. For households holding a disproportionate amount of money, their welfare is clearly negatively related to inﬂation. In contrast, households holding relatively less money can realize welfare gains from higher inﬂation.

Asset Prices: Apart from heterogeneous cash holdings, it appears that the inequality effect from monetary policy also runs through asset prices.

Through two papers, Kaplan, Moll and Violante (2016a) and (2016b) suggest that monetary expansion does not directly induce consumption. Most of the boost to aggregate consumption that takes place comes about very indirectly: widening spreads increasing the attractiveness of illiquid financial assets which in turn trigger an investment boom, causing a multiplier effect on aggregate demand and only eventually feeding into increases in labor income and thereby consumption. Kaplan Moll and Violante imply that this highly indirect mechanism for generating growth and consumption leaves increased inequality in its wake. Indeed, to solve their model, Kaplan Moll and Violante need to assume a leptokurtic distribution of annual earnings, consistent with empirical evidence produced by others. Saiki and Frost (2014) see direct empirical evidence of the portfolio effect on inequality in Japan: an increase to the monetary base through purchases of both safe and risky assets tends to increase overall asset prices, primarily benefiting upper incomes. While Saiki and Frost found this unconventional monetary policy finally “bearing fruit” in the years after the Great Recession, this effectiveness came with the unwanted side effect of increased inequality. The authors speculate that “the portfolio channel will be even larger in the US, UK, and many Eurozone economies, where households hold a larger portion of their savings in equities and bonds”.

In summary, monetary expansion exacerbates inequality when it works through the Portfolio Channel. Expansionary monetary policy reduces the “return” on cash which hurts poorer households who hold disproportionate amounts of cash. Also, monetary expansion’s effects on consumption are likely very indirect, working first on raising asset prices to trigger investment that only eventually feeds back into labor income but meanwhile leaving increased inequality through benefiting asset owners.

Conclusion: We saw above that the circumstantial and statistical evidence is highly suggestive of monetary stimulus having adverse redistributive impact. The academic literature suggests expansionary monetary policy works through five channels, four of which unambiguously increase inequality. A fifth channel, which generally reduces inequality, can under some circumstances reverse course and further increase inequality. Only detailed empirical analysis can confirm the strength of each of the five channels and their overall net effect. However, pulling all the existing theory and evidence together, it seems reasonable to argue that monetary liquidity exacerbates inequality

Inequality Encourages Charity

The previous section lays out theoretical and empirical arguments that liquidity exacerbates inequality. But what does that say about charity? Does greater inequality lead to greater charitable donations?

To answer the question, we draw on the literature on collective action and voluntary provision of public goods that started with Bergstrom, Blume and Varian (1986). Drawing on this literature, Karaivanov (2008) develops a game-theoretic model where households have different levels of wealth and propensity for charity (clearly a kind of “public good”). The insight from his model is that total charitable donation depends on the average levels of wealth and propensity towards charity, as well as the correlation between the two. In general, inequality can alleviate the free riding problem since it introduces the possibility for households with high propensity for charity or high endowments to contribute regardless of the action of the others (the model shows one exception to this — in the case where not only does the production function of charity have increasing returns to scale but households have both high average wealth and high average propensity for charity. However, whatever the shape of the production function, we know that society has neither high average wealth nor high average propensity for charity).

Another interesting insight from Karaivanov’s model is that the highest equilibrium contribution occurs when the contributors have both high wealth and propensity to contribute which are positively correlated. This implies that if it is possible to select among all potential contributors those with most favorable characteristics, one can achieve higher total contribution level per given number of people.

Karaivanov’s model fits well with empirical observation. Higher wealth inequality seems to drive higher charitable donations. Total donations also seem responsive to situations where the propensity for charity rises e.g. in hard times or when people seem to make more use of the services of charities and endowments. Total donations also seem to rise when targeted campaigns induce higher levels of propensity among the wealthiest, exactly as Karaivanov suggested.

Donations Are Driven by the Distribution of Income and Wealth

There appears to abundant empirical evidence to support the Karaivanov (2008) model of wealth driving charitable donations.

Wiepking and Bekkers (2012) reviewed the literature on charitable giving and found “overwhelming evidence for a positive relationship between income, wealth and amount donated to charitable organizations” and also that “level of wealth has a stronger relationship with the level of giving than the level of income from employment”. According to List (2011), around 75% of charitable donations come from individuals and around 90% of households with incomes above $130,000 give to charity; such households give on average $4,500 annually to charity. Households in the top 4% of the income distribution were estimated to give over 40% of total charitable contributions in 1995.

Brown Dimmock and Weisbenner (2012) found that charitable donations to higher educational institutions are positively and significantly related to per capita income, the returns of local stocks and house values. The relationship was stronger for capital donations i.e. for longer term, more restricted use of funds. When donors are doing better financially, they donate more to higher education.

Wiepking and Bekkers (2012)’s review of the literature also covered how responsive donations are to income. Their summary of the existing research is that “giving is comparable to basic good consumption and inelastic: people increase donations when gaining income, but the increase in giving is lower than the increase in income”. Also, not only is the responsiveness of giving inelastic to income, it also appears to decline with income. Wiepking and Bekkers (2012) quote a study by Bakija and Heim (2008) that finds “(persistent) income elasticities of 0.846 for households with an income below U.S.$200K, 0.806 for incomes between U.S.$200K and U.S.$500K, 0.699 for incomes between U.S.$500K and U.S.$1M, and 0.659 for incomes over U.S.$1M”.

Although donations seem to be positively related but not elastic to income, there are important nuances to this broad pattern, starkly brought out by a study by Hughes Luksetich (2008) that tracked couples from 1994 to 2001. First, donations appear to be highly elastic to increases in permanent income. The authors found that a 10% increase in permanent family income raised charitable donations by 17.5%. A second interesting and somewhat counter-intuitive nuance is that personal indebtedness has a strong positive impact on charitable donations. Indeed, Hughes Luksetich (2008) found that the level of personal indebtedness is significantly more impactful than wealth itself. This may well be related to the asymmetric behavior of households at their borrowing limit uncovered by Auclert (2016) : they have to cut consumption one for one when squeezed but have low marginal propensity to consume from modest loosening of their budget constraints.

Another question is how changes in inequality impact donations. Payne and Smith (2015) evaluate this issue using Canadian data between 1991 and 2006. Over this period, real donations per household (2005 dollars) increased significantly across the sample period from roughly $369 to $649 (76%) even as, tellingly, the fraction of households donating fell from 52% to 44%. Measured by Gini coefficient, the (smaller area) postal sorting area-level Gini rose by about 16% and the (larger area) city-level Gini rose by 13%. Based on their regressions, for the 76% rise in donations, the authors attribute 3.5% of the growth in donations to rise in inequality in the immediate postal sorting area, and 8.2% to rise in inequality at the city level. The regressions controlled for various factors including income levels but also education, ethnic diversity, family structure, home values, age distribution and labor market conditions. However, the authors also found that changes in inequality had less effect on donations in where inequality was already high at both local and wider neighborhood levels.

Donations Are Also Influenced by Propensity for Charity

We saw above that Karaivanov‘s (2008) theoretical model suggests that donations respond to rising propensity for charity. Resonating with this, Payne and Smith (2015) also noticed in their research on Canada that not only did increasing inequality contribute to rising donations, but the share of gifts to welfare, education and health increased at the expense of religious charities. The authors speculate that people may respond to rising inequality by donating more to redistributive charitable goods.

More direct evidence of the role of propensity is also available. Fund-raising campaigns by foundation and endowments are direct instruments to induce greater propensity — and they appear to work well. Okten and Weisbrood (2000)’s analysis of US tax returns from 1982 to 1994 showed that fundraising expenditures appeared to increase the overall levels of donations. The informational benefits of fundraising expenditures appeared to more than offset the money diverted from programs to fundraising. A variety of other studies have also shown that donors respond positively to fundraising expenditures at nonprofits. Andreoni (1998) emphasized the role that fund-raising campaigns played in donations, with seed grants or gifts from “leadership givers” being pivotal in inducing others to join the donations bandwagon. An example of such a fund-raising drive was the state governor’s 1995 offer of $27 million in state bonds followed by an additional $25 million contribution by a wealthy senator for a new basketball arena for the University of Wisconsin. These leadership gifts triggered other donations leading up to a total of $72 million raised. It is no wonder then that, in 1995, the 25 largest charities spent an average of over $25 million each on fund-raising: this amounted to as much as 14 percent of total charitable gifts.

Donor propensity also appears to rise when they observe growing demand for their foundation or endowment’s services. Skinner, Ekelund and Jackson (2009) found that attendance at museums (a proxy for demand for museum services) between 1990 and 2000 ran in a direction opposite to the business cycle while funding to museums ran in the same direction as the business cycle. However, using statistical analysis, it appeared that museum attendance in fact seemed to “cause” federal contributions to museums. Such federal contributions, through their implied certification of quality, induced other government and private contributions.

To be sure, diminishing propensity to donate can sometimes undermine overall charitable donations. Meer, Miller and Wulfsberg (2016) studied empirical patterns of giving before, during and after the Great Recession. They found sharp declines in overall donative behavior during the Great Recession and after, not accounted for by shocks to income or wealth. Of course, the magnitude of the Great Recession was such that it may have had some fundamental effects on propensity to donate in bad times — time will tell how lasting these effects have been and what kind of moderating impact they will have on overall donations. However, when placed within Karaivanov’s framework — donations being driven by both average wealth and propensity to donate — we can see that higher inequality caused by the Great Recession could plausibly have outweighed the decline in propensity to donate.

The Hoarding of Donations

Endowments seem to behave like donors to themselves, through asymmetric adjustments to their payout ratios. When times are bad, they “donate” more to themselves by reducing payout ratios to conserve the endowment. But they do not raise payouts in good times.

A study of 1987–2009 panel data by Brown, Dimmock, Kang and Weisbenner (2014) found that: “university endowments respond asymmetrically to contemporaneous positive and negative financial shocks. In response to contemporaneous positive shocks, endowments tend to leave current payouts unchanged. Such behavior is consistent with endowments following their stated payout policies, which are based on past endowment values and not current returns, in order to smooth payouts (e.g., pay out 5 percent of the past three-year average of endowment values). However, following contemporaneous negative shocks, endowments actively reduce payout rates. Unlike their response to positive shocks, this behavior is inconsistent with endowments following their standard smoothing rules. This asymmetry in the response to positive and negative shocks is especially strong if we explicitly control for the payout rate that is implied by the universities’ stated payout rules (something we do for a subsample of the endowments for which we have sufficient information to precisely document their payout rules). We also fail to find consistent evidence that universities change endowment payouts to offset shocks to other sources of university revenues. These findings, which we confirm through several robustness checks, suggest that endowments’ behavior differs from that predicted by several normative models of endowment behavior. Instead, our results support an alternative hypothesis, which we refer to as ‘endowment hoarding’”.

Given the observed pattern of endowment hoarding in response to financial shocks, it should be no surprise that a lowering of interest rates (signaling a lower return environment) would drive endowment managers to reduce payout ratios, without any pressure to reverse course when interest rates rise. A glimpse of this thinking is visible in the comments from Larry Summers, the former president of Harvard: “If it makes sense for Harvard University to pay out 5 percent of its endowment in 1999 when the real interest rate was 4 percent, it’s really quite unlikely that it makes sense to pay out 5 percent of its endowment in 2016 when the real interest rate is zero….it’s hard to believe that the real rate should have moved by 4 percent and there should be no change in the types of spending rules that people operate”.

Conclusion: We saw in this section that charity is driven by the combined effect of average wealth level and propensity to donate. There is enough empirical evidence that charitable giving is positively but inelastically related to income and wealth. Furthermore, the worsening of inequality triggers even more charity. Propensity to donate also plays an important role in charity. Usually (but not always), propensity to donate is sensitive to need e.g. when times are bad or when charitable services are more intensively required. Charitable endowments seem to respond in similar fashion, displaying higher propensity to donate to themselves through hoarding when times are bad. Overall, given the theoretical and empirical evidence, it is clear that inequality stimulates charity.

Charity Necessitates Temerity

Liquidity exacerbates inequality which encourages charity. How does that in any way relate to risk-taking?

Steady Donations Create a Synthetic Hybrid Bond

As we have seen from previous sections, liquidity-exacerbated inequality encourages charitable donations through good times and bad. Accordingly, after adjusting for payouts, charities keep receiving annuity-like donations; however, they do see some ups and downs in line with equity / real estate fluctuations.

In effect, donations have the profile of a near-term bond with an equity kicker. Mostly annuity inflows (after adjusting for payouts) but garnished with ups and downs. The Blanchett (2014) study of donation volatility corroborates this reading. He regressed annual changes in donation against the Fama & French 5 risk factors, also adding Momentum and Liquidity risk factors: i.e. market return risk, size risk (return on small cap minus big cap stocks), value risk(return on value minus growth stocks), bond duration risk (return on long government bond less short term treasury), bond default risk (return on long corporate bond less long government bond), momentum risk (stocks that have recently done well continue to outperform) and liquidity risk (illiquid securities outperform liquid securities). His regression results have relatively good explanatory power (R-squared = 53–54%) for both individual and total donations. All factors are found to be very significant (significant at 1% level) except that size risk and liquidity risk are not significant (even at 10% level). The relationship is negative for bond duration and bond default risks but positive for all others.

The signs and significance levels of the respective coefficients suggest that donation risk looks mainly similar to that of a short-term bond and also somewhat like a high-book value momentum chasing stock. It hardly bears hardly any resemblance to the risk profile of small-cap or illiquid stocks.

One way of looking at Blanchett’s findings is that for an endowment, donation revenues behave like a low risk, short term bond with a small equity component attached.

Donations as Synthetic Hybrid Bonds: Wealth Effects and Optimizing the Portfolio

Carrying the analogy forward, the “value” of such a “hybrid bond” would then vary inversely with interest rates but also vary on a lagged basis with ups and downs in the equity market.

If donation inflows create a synthetic hybrid bond, what happens to the wealth of the “bondholders” when interest rates fall? And how should the “bondholders” manage their portfolios when interest rates fall? The bond like profile of payout-adjusted donation revenues described above means that the “value of the bond” rises when interest rates fall. Thus, the endowment (the bondholder in this instance) suddenly feels wealthier on the “asset” side. Moreover, not only would the value of this “hybrid bond” jump with cuts in interest rates, holders of this “hybrid bond” would feel that the value of their asset portfolio has jumped with such rate cuts. Stretching the analogy even further, just like many individuals who rely on ongoing cash streams feel insulated from changes in the capital value of their bond holdings, endowments do not automatically feel or behave poorer when interest rates rise.

If donation streams are accurately described as a synthetic hybrid bond, what does that imply for how the “bondholders” i.e. the endowments should manage the rest of their portfolio? Blanchett (2014) runs simulations to model the optimal portfolio. His findings are consistent with generally understood principles of portfolio allocation. Equity allocations ought to be higher when donation revenue is a greater portion of overall revenue of the endowment. Also, donations that are more stock-like ought to result in lower equity allocations for a given level of risk aversion. Finally, equity weights should decline as the percentage of income derived from the donations decreases.

Rosen and Sappington (2015) carry out a study of public and private university endowments between 1987–2009 to see how responsive they are to all university income inflows — tuition, public funding and private donations. They find that universities that expect higher levels of these inflows (i) are more likely to invest in alternative assets; and (ii) allocate a larger percentage of their endowments to alternative assets. The likelihood that a university decides to include alternative assets in its portfolio in a given year increases by 11.3 percentage points with a one standard deviation increase in expected income, and decreases by 8.2 percentage points with a one standard deviation increase in the variability of income. Importantly, the share of a university’s portfolio held in alternative assets increases by 7.5 percentage points and decreases by 1.1 percentage points with a one standard deviation increase in expected background income and background risk, respectively.

Rosen and Sappington (2015)’s findings confirm and complement earlier work by Dimmock (2012) who studied university data in the 2002–2003 academic year and examined the effect of variability of university income on asset allocation by their endowments. Dimmock (2012) found that a one standard deviation variation in total university inflows caused a 0.6% points inverse variation in endowment portfolio volatility. What is more, variability in university inflows had considerably more impact on alternative assets such as venture capital, private equity and hedge funds than on equity, which the author attributes to the larger standard deviations (or higher moment risk in case of hedge funds) of such alternative assets. A one standard deviation variation in total university inflows caused an inverse variation of 3.2% points in alternate assets. Among a host of university endowment indicators considered in the regression, average donation is the most significant variable explaining asset allocation.

Conclusion: Donations, stimulated by liquidity-exacerbated inequality, create a stream of inflows to charitable endowments, modulated slightly by ups and downs in equity and real estate markets but otherwise steady and regular. These flows have the characteristic of a bond with equity kicker, as statistical tests confirm. Endowments, who are “holders” of this “bond”, diversify and balance out their investment portfolio by investing in assets whose risk characteristics offset those of the bond. In particular, they invest in alternative assets such as private equity, venture capital and hedge funds.

Concluding Summary

By peeling back several layers of the onion, we find the route through which monetary liquidity gets translated into higher risk-taking by charitable endowments. The mechanisms are complex and work through multiple channels and over several steps.

Monetary expansion (“Liquidity”) worsens the distribution of income, consumption and wealth (“Inequality”). This happens through five channels — (1) Income Composition Channel: profits gaining at the expense of wages (2) Financial Segmentation Channel: those closest to the financial system gaining at the expense of those further away (3) Borrowers versus Savers Redistribution Channel: asset-holders gaining through interest rate cuts while borrowers find their liabilities revalued upward (4) Earnings Heterogeneity Channel: those at the upper end of incomes could benefit disproportionately if monetary expansion happens under a less than progressive tax and unemployment benefits system (5) Portfolio Channel: Expanding the supply of money reduces its return, hurting the poor who tend to largely hold their little wealth in the form of cash even as the purchase of both safe and risky assets enabled by the expanded monetary base benefits rich asset owners.

With increased Inequality from Liquidity via these five channels, there is a clear impact on the overall level of charitable donations (“Charity”). High income households that have high propensity for charitable action find more resources available to do so and also tend to care less if other households do not contribute their share. They find it possible to contribute even more if there is more demand for the charitable activities they support e.g. when times are hard. The empirical evidence seems to suggest that wealthy donors increase their charitable donations with higher income and wealth. But charitable donations depend not only on redistribution effects experienced say through increases in equity prices or real estate prices but also on the demand for donations from endowments themselves. Charitable endowments are found to actively court donors when need or demand for their services is high; indeed, the endowments also tend to themselves hoard more at such times. As an outcome of all of the above, endowments keep receiving regular donations through good times as well as bad; however, they do see some ups and downs in line with equity / real estate fluctuations. This pattern of flows can easily be recognized as similar to an annuity income with an upside for equity and real-estate prices; in other words, some kind of “synthetic bond with an equity kicker”.

From the above, we see that Liquidity leads to Inequality which in turn encourages Charity. The “synthetic bond with equity kicker” nature of the charitable donations flow have their own implications. To balance out the low-risk nature of the “synthetic bond with equity kicker”, holders of the bond must take more active risk. They do this through investing in high-risk asset classes such as venture capital and hedge funds (“Temerity”).

Although the logic is clear, we must be cautious that is as yet only a working hypothesis. In an effort to establish the bigger picture, we have stitched several mechanisms, channels and impulses together. Several impulses could conceivably have secondary or tertiary effects or feedback loops that have not been considered here. Also, we have yet to quantify the strength of the different channels; and we are not sure whether they are additive or if some may offset each other. Finally, a careful analysis must navigate minefields of endogeneity and possible autocorrelation, given the large number of factors considered. Given all of the foregoing, a systematic theory will perhaps require general equilibrium modeling and should also be calibrated against real world information. That is outside the scope of our work here. Instead, this paper attempts to describe the relevant channels and transmission mechanisms discussed in the literature. We offer a qualitative synthesis as a working hypothesis that seems to fit well with the anecdotal wisdom of market practitioners and also with existing empirical evidence and the historical record.

References

Albanesi (2007) “Redistribution and Optimal Monetary Policy: Results and Open Questions”, Rivista di Politica Economica, July-August

Andrei and Ledoit (2015) “The Redistributive Effects of Monetary Policy”

Andreoni (1998) Toward a Theory of Charitable Fund-Raising, Journal of Political Economy, Vol. 106 №6 (December)

Auclert (2016) “Monetary Policy and the Redistribution Channel”, January

Balac (2008) “Monetary Inflation’s Effect on Wealth Inequality”, Quarterly Journal of Austrian Economics, 11:1–17

Bergstrom, Blume and Varian (1986) “On the Private Provision of Public Goods”, Journal of Public Economics 29 North-Holland

Bernanke (2015) “Monetary Policy and Inequality”, Brookings Institution Blogs, June 1

Bezemer (2014) “Schumpeter might be right again: the functional differentiation of credit”, Journal of Evolutionary Economics, Volume 24, Issue 5 (November)

Bhattacharya Haslag and Martin (2004) “Heterogeneity, Redistribution and the Friedman Rule”, RWP 04–01 Research Division Federal Reserve Bank of Kansas City

Bivens (2015) “Gauging the Impact of the Fed on Inequality During the Great Recession”, Hutchins Center on Fiscal & Monetary Policy at Brookings, Working Paper #12, June 1

Bloomberg (2016) “Ex-Harvard Head Summers Suggests School Curb Endowment Payout”, April 19

Brown, Dimmock, Kang and Weisbenner (2014) “How University Endowments Respond to Financial Market Shocks: Evidence and Implications”, American Economic Review, 104 (3)

Blanchett (2014) “Donation Risk and Optimal Endowment Portfolio Allocations”, Journal of Portfolio Management, Fall

Brown, Dimmock and Weisbenner (2012) “The Supply of and Demand for Charitable Donations to Higher Education”, NBER Working Paper 18389

Cohan (2014) “How Quantitative Easing Contributed to the Nation’s Inequality Problem”, The New York Times, October 22

Coibion, Gorodnichenko, Kueng and Silvia (2012) “Innocent Bystanders? Monetary Policy and Inequality in the US”, NBER Working Paper 18170

Coibion, Gorodnichenko, Kueng and Silvia (2014) “Innocent Bystanders? Monetary Policy and Inequality in the US”, VOX CEPR’s Policy Portal, October 25

Dimmock (2012) “Background Risk and University Endowment Funds”, Review of Economics and Statistics, August, 94(3)

Domanski Scatigna and Zabai (2016) “Wealth inequality and monetary policy”, BIS Quarterly Review, March

Gornemann Kuester and Nakajima (2016) “Doves for the Rich, Hawks for the Poor? Distributional Consequences of Monetary Policy”,

Heathcote Perri and Violante (2009) “Unequal We Stand: An Empirical Analysis of Economic Inequality in the United States, 1967–2006”, NBER Working Paper No 15483 November

Hughes and Luksetich (2008) “Income Volatility and Wealth: The Effect on Charitable Giving”, Nonprofit and Voluntary Sector Quarterly, Issue 37

Kaplan, Moll and Violante (2016) “Monetary Policy According to HANK”, NBER Working Paper 21897, January

Kaplan, Moll and Violante (2016b) “A Note on Unconventional Monetary Policy in HANK”, January 28 Discussion Paper

Karaivanov (2008) “Heterogeneity, Returns to Scale and Collective Action”, May

Krueger, Mitman and Perri (2016) “Macroeconomics and Household Heterogeneity”, NBER Working Paper 22319, June

List (2011) “The Market for Charitable Giving”, Journal of Economic Perspectives, Vol 25 No 2, Spring

Meer, Miller and Wulfsberg (2016) “The Great Recession and Charitable Giving” , NBER Working Paper 22902

Okten and Weisbrod (2000) Determinants of donations in private nonprofit markets, Journal of Public Economics 75(2000)

Payne and Smith (2015) “Does Income Inequality Increase Charitable Giving?”, CLSRN Working Paper №150, January

Rosen and Sappington (2015) “What Do University Endowment Managers Worry About? An Analysis of Alternative Asset Investments and Background Income”, NBER Working Paper 21271

Saiki and Frost (2014) “How Does Unconventional Monetary Policy Affect Inequality? Evidence from Japan”, DNB Working Paper №423, De Nederlandsche Bank NV Amsterdam The Netherlands, May

Skinner, Ekelund and Jackson (2009) Art Museum Attendance, Public Funding and the Business Cycle, American Journal of Economics and Sociology, Vol. 68 №2 (April)

Watkins (2014) “Quantitative Easing as a Means of Reducing Unemployment: A New Version of Trickle-Down Economics”, paper presented at the 2014 Annual Meetings Association of Evolutionary Economics, Philadelphia, January 4

Wiepking and Bekkers (2012) “Who Gives? A Literature Review of Predictors of Charitable Giving II, Gender Family Composition and Income”, April 12

Williamson (2005) “Monetary Policy and Distribution”, http://repec.org/sed2005/up.2424.1106844165.pdf

Liquidity, Inequality, Charity and Temerity

Written by Viren Mahurkar