According to one theory, consumer-voters mistake lump-sum aid for matching aid such that lump-sum aid reduces their tax price of the aided service. Consumer-voters misunderstand that the substitution effect of matching aid comes from lump-sum aid, resulting in higher expenditures for the aided service. This flypaper effect is more likely when both lump-sum aid and matching aid (or equivalent tax-price reducing mechanisms) coexist and there is a high likelihood of interaction between the lump-sum aid and the substitution effect. However, surprisingly few studies have developed a clear formula to evaluate the flypaper effect while the literature generally assumes the interaction as one of potential causes for the flypaper effect. This paper fills in the huge gap in the literature by providing a formula to show whether and how lump-sum aid causes the flypaper effect.
Keywords: flypaper effect, substitution effect, matching aid, lump-sum aid, fiscal illusion
Numerous theoretical and empirical studies have reported that a one-dollar increase in lump-sum aid has much stronger fiscal impacts on recipient governments' expenditures on the aided service than an equal amount of increase in consumer-voters' income does. This phenomenon has been known as the flypaper effect. One dominant theory explaining the flypaper effect is the theory of fiscal illusion on the part of consumer-voters in the recipient governments. Consumer-voters mistake the lump-sum aid for matching aid and might perceive that the lump-sum aid reduces their marginal tax price of the governmental service aided by donor governments' lump-sum aid. The lower tax price generates the substitution effect and consumer-voters tend to demand the aided service more. As a result, the expenditure for the service grows as well. This possibility is more likely whenever lump-sum aid and matching aid (or similar fiscal settings that render tax-price reducing impacts from matching aid) coexist and the two types of aid interact. However, surprisingly few studies provide a clear formula to measure the fiscal impact of the misperceived tax price on governmental expenditures on the aided service. This paper develops an explicit formula to measure the fiscal impact (i.e., the substitution effect), which interacts with lump-sum aid, by using Ohio's local school district data. The formula to measure the interacted substitution effect can be easily and clearly applied to similar aid structures in other states.
The next section introduces the literature on the flypaper effect. The following section explains a diagrammatic measurement of the substitution effect. The next section succinctly presents a model of expenditure for Ohio's local school districts. After that, the section on data and methodology explains data sources and model estimation strategies. Empirical findings in this paper clarify that the substitution effect works through the lump-sum aid as the theory of fiscal illusion predicted. In addition, the findings clearly reveal the route, through which matching share, typically associated with matching aid, influences school district expenditures via lump-sum aid.
THE FLYPAPER EFFECT AND NEW MEASUREMENTS
2.1 THE FLYPAPER EFFECT IN THE LITERATURE
Fiscal impacts of lump-sum aid have been observed to be stronger than those from the increase in consumer-voters' personal income. A one-dollar increase in the latter typically raises the expenditure on public goods delivered by governments receiving the lump-sum aid by about $0.05 to $0.10 (Cullis and Jones 2009, 388). In contrast, a one-dollar increase in the lump-sum aid tends to increase the public expenditure by about $0.25 to $0.50. This phenomenon has been known as the flypaper effect (Gramlich and Galper 1973; Fisher 2016, 229; Hines and Thaler 1995; Hyman 2011, 732-733; Rosen and Gayer 2010, 532-533). Some other studies report slightly stronger flypaper effect up to about one dollar (Case et al. 1993; Gamkhar and Oates 1996; Gennari and Messina 2013).
One of the most dominant theories explaining the flypaper effect is the theory of fiscal illusion either by consumer-voters or budget-maximizing bureaucrats. Consumer-voters might perceive that lump-sum aid reduces marginal cost of service deliveries by the governments receiving the aid. If the governments spend $100 per consumer-voter for a certain service and receive a $20 per consumer-voter lump-sum aid for the spending, the recipient governments shoulder only 80 % of the cost incurred for the service. Consumer-voters are likely to mistake the reduction in the service delivery cost for a decrease in their tax price. The final fiscal impacts might be similar to the effect from matching grants that reduce tax prices through matching share (Courant, Gramlich, and Rubinfeld 1979; Oates 1979). If matching share is associated with local tax payment, as in the case of Ohio's property tax rollback credit for local property taxpayers, this type of fiscal illusion would also be highly probable, as this paper shows.
In a similar vein, budget-maximizing bureaucrats also fall in fiscal illusion, but in this case over average costs of the service delivery. King (1984) applied the framework in Niskanen's (1971) budget-maximizing bureaucrats to explain the flypaper effect. Budget-maximizing bureaucrats tend to request budgets larger than socially efficient ones where marginal benefits from the governmental service equal marginal costs of the service delivery. When their governments receive lump-sum aid, they misunderstand that the lump-sum aid lowers the entire costs of the service delivery and ultimately reduces average costs of the service delivery. This makes a downward shift of the cost curve in the Niskanen model. As a result, the lump-sum aid raises the governmental budgets since bureaucrats pursue budget-maximization on the newly perceived average costs. Some empirical findings support this possibility of fiscal illusion on the part of governmental bureaucrats. When bureaucrats in the governments receiving the lump-sum aid kept stronger budget-manipulating power, the flypaper effect was much higher (Bae and Feiock 2004; Schneider and Ji 1987). Budget-maximizing bureaucrats might also control the information on the lump-sum aid and induce consumer-voters to support higher spending while at the same time, they spend the lump-sum aid. This might also explain why the lump-sum aid can significantly increase the expenditures for the public services that receive grants from donor governments (Filimon, Romer, and Rosenthal 1982).
Other competing theories pay attention to the grant-decision process. When donor governments choose recipient governments for the lump-sum aid, they are likely to distribute their lump-sum aid to recipient governments that are willing and likely to spend larger amounts of local dollars. This process turns the lump-sum aid into a de facto matching aid with much higher fiscal impacts (Chernick 1979). Inman (2008) especially points at the political aspect inherent in the grant-decision process by citing Knight (2002). The legislators whose districts benefit most highly might be willing to spend more on the aided governmental service and as a result, tend to make the winning bids for the lump-sum aid. The end-result is a positive correlation between the lump-sum aid awarded and recipient governments' spending.
Another provocative recent study indicated that the analytical methodology employed in the previous studies might have overestimated the flypaper effect. When logarithmic model specification was used, instead of linear specifications between the lump-sum aid and recipient governments' spending, the flypaper effect was much weaker (Grizzle 2011). However, more recent and elaborate studies strongly confirm the flypaper effect noted above. For instance, Roemer and Silvestre (2002) indicate the single-policy-dimension and median voter framework in the previous studies. They allow for more complicated party competition under multidimensional issue frameworks but still support the previous studies. Sobel and Crowly (2014) investigate how an inter-temporal dimension of recipient governments' spending and donor governments' grants affects the flypaper effect. Since governmental programs, which might have expanded by grants, are hard to cut, future recipient governments' budgets, especially tax revenues, tend to grow. Their empirical findings show that grants tend to increase future state and local tax increases by about $0.40.
2.2 NEW MEASUREMENTS OF THE FLYPAPER EFFECT
Among the various theories explaining the flypaper effect, the theory of fiscal illusion by consumer-voters has recently garnered an interesting attraction. Rockoff (2010) developed an elaborate formula to measure consumer-voters' fiscal illusion. However, his formula attempted to tap how much lump-sum aid dampens the price effect of matching aid rather than how much matching aid affects the fiscal impact of lump-sum aid. Based on his formula, his estimated flypaper effect in New York's lump-sum aid to school districts is about $0.1, which is much lower than the range of the flypaper effect introduced above. One potential reason for this low estimation of the flypaper is because his interaction formula is developed primarily to tap how lump-sum aid affects matching aid, not the other way around. Duncombe and Yinger (1998, 2009; see also Eom et al. 2014) have developed one of the most elaborate measures of the interaction between lump-sum aid and matching aid, which is defined as DY Interaction in this paper. DY Interaction is a theoretically elaborate measure because unlike Rockoff's interaction measure, it is derived from consumer-voters' and local governments' fiscal optimization equations. However, DY Interaction does not specifically test whether consumer-voters' fiscal illusion explains the flypaper effect.
MATCHING AID, LUMP-SUM AID, AND SUBSTITUTION...